Month: April 2025

Lumentum at Inflection Point with 20% QoQ Growth in AI-Related Segment

Posted on by io-fund

Lumentum has been on our radar for some time as the company supplies components for datacom transceivers and optical interconnects. Lumentum is a small-cap company with differentiated technology that has caught the attention of heavyweight NVIDIA. We’ve been closely monitoring Lumentum for roughly a year, waiting patiently for their EML lasers for 200G to ship, enabling 800G and 1.6T bandwidths. Any progress here should continue into 2026-2027 for 400G data lanes and 3.2T bandwidths. 

Optical interconnects help data centers accelerate data throughput between data centers and inside the data center between servers or racks, while reducing latency and power consumption. AI is driving cloud demand higher from the hyperscalers, leading to more data being created and processed, thus helping drive a need for these interconnects to meet demand for high-speed, low power data transmission in data centers. 

The company supplies three types of lasers for datacom transceivers: VSCELs, CW lasers for silicon photonics and EML-based lasers. Per our previous analysis, the 200G EMLs are what is expected to drive an inflection in 2025.  

In the most recent quarter, Lumentum reported their Cloud and Networking segment grew 18.3% YoY and 20.2% QoQ to $339.2 million. This caused total revenue to grow 9.7% yet total revenue is expected to accelerate to 46.6% and 43% in the second half of the year.  

More on Lumentum’s EML Lasers: 

EMLs were traditionally used by telecom customers, yet became attractive for AI servers due to meeting the 200G per second speeds necessary for 1.6T optical modules to support AI models. These are called single mode optics, made of Indium Phosphide, which has been used instead of silicon for long-haul networking due to being a superior choice for optical functions, such as enabling the laser, modulator, photodetector and amplifier.  

InP is more expensive at the component level as four EMLs are needed compared to two lower-cost CW lasers for silicon photonics modules, yet this difference at the component level can be made up for in data centers as InP reduces power consumption. 

In the December quarter, Lumentum stated they began shipping its 200G lane speed externally-modulated lasers (EMLs) to multiple customers. The Company stated they would increase its EML capacity by 40% YoY from June 2024 to June 2025, followed by another 40% increase by the end of calendar 2025 as its Thailand production accelerates – representing compounding growth in an 18-month time frame:   

“So, we're just overall the demand is outstripping even that 40% compounded twice in an 18 month period of time. We do have additional capacity expansion beyond the end of calendar '25 obviously and those are investments that we've made over the past several quarters that come online in calendar '25, which will give us increased capacity as well.” 

Lumentum’s indium phosphide 100G EMLs (Externally-Modulated Lasers) are being shipped and used in leading single-mode 400G and 800G optical transceivers. These customers are now designing the company’s 200G EMLs into their next generation of transceivers. Already being a lead supplier for 100G EML transceivers, the company is set up to be in pole position for the 200G EML transceivers. 

Per the most recent earnings call: “We achieved another record for EML unit shipments in Q2 and began delivering 200G lane speed EMLs to multiple customers.  

Based on the breadth of our 200G EML design wins, we expect to gain additional laser transmitter market share in the upcoming wave of 800G and 1.6T transceivers, utilizing the more efficient 200G EMLs for AI applications. Complementing our EMLs are our new 200G lens integrated photodetector arrays, which adds to our content opportunity in next generation 800G and 1.6T transceivers as well as strengthens our vertical integration strategy for our own cloud modules.” 

Specifically, the company’s experience in InP long haul transceivers is being tapped as AI servers scale, especially since InP reduces power consumption compared to silicon. Lumentum is also working on higher speed optical links, including 400G per lane. This means more data, fewer lanes if you compare it to 200G per lane. This is important especially as data centers will eventually move not only to 1.6 terabits per second (1.6T) (4X400G or 8X200G) but also 3.2T (8X400G).  

Management also stated the new 200G lens integrated photodetector arrays “adds to our content opportunity in next generation 800G and 1.6T transceivers.” 

Lumentum’s Co-Packaged Optics Opportunity for High Power Lasers:  

Silicon photonics are the only viable choice for rack-to-rack and across the data center due to the need for high bandwidth and lower power at high speeds. There is also low-loss over long distances with optical fiber, which refers to preserving the original signal, whereas copper sees signal degradation over longer distances.  

Where there is a debate and an important shift occurring is in the networking between chips and inside the rack. We’ve recently covered the benefits of co-packaged optics as a replacement for pluggable transceivers when networking between chips. At 800G and 1.6T speeds, the electrical signals over power circuit boards (PCBs) run too hot, are power hungry and create loss for the signal. The overall goal is to move the optics closer in proximity to overcome scaling issues when increasing electrical speed. 

Nvidia stated that by replacing pluggable optics with silicon photonics on the package, it can “deliver 3.5x more power efficiency, 63x greater signal integrity, 10x better network resiliency at scale and 1.3x faster deployment compared with traditional methods.”  

Nvidia VP Ian Buck stated at GTC that the CPO switches help reduce power consumption by eliminating the need for external lasers and pluggable transceivers to achieve a significant reduction in power from 39 watts to 9 watts. Buck explained that this “gives you that benefit from going from 39 watts of power down to only 9 watts of power for the same number of ports, and that's huge. It doesn't sound like 39 sounds a lot. But if you get 400,000 GPUs in an AI supercomputer, there's like 24 megawatts of lasers like so that's a lot of laser light that could be optimized and made more efficient.” 

While Lumentum may see some decline in its pluggable optics transceiver business as the industry shifts, it is also well-positioned to benefit from the transition. The company is a key supplier of high-power lasers required for co-packaged optics (CPO), which could offset near-term losses. Taken together, the impact should be at the onset net neutral — with (potentially significant) upside over time from growing demand for EML lasers, the transition to 200G and 400G per lane, plus optical switching (see below). 

Regarding CPOs, here is what was stated on the most recent call: “But I'd also highlight for Lumentum a little bit of a unique situation where we're more modest share on transceivers today and share gaining, if you will, over time. So even if there is some cannibalization in the mid to long-term, we don't think it impacts our transceiver opportunity and creates an expanded opportunity for high power lasers.” 

Nearly a year ago, Lumentum announced an ultra-high output power 1310 nm DFB laser in beta, stating it was designed to reduce the “required number of lasers” while boosting efficiency and reliability in large scale AI/ML infrastructure. 

It was also shared that Lumentum is shipping preproduction volumes now related to CPOs: 

 ‘As part of one collaboration, we began shipping preproduction volumes of our unique ultra-high power lasers to an AI infrastructure customer for a proprietary interconnect solution in Q2 and have received follow-on orders as well as excellent feedback on the product's performance. This is a very exciting opportunity.” 

It was later teased out on the earnings call that 2026 could be a big year for Lumentum’s high power laser: “I would say preproduction volume with forecasts for more meaningful volume throughout calendar '25 and then real extremely high volume in calendar '26. So, feedback is very positive. The performance is very unique and the line width that we produce with the laser is superior to anything in the market.”

CPO (with Switch ASIC) + Optical Switch + GPU Clusters = The Future of Scale Up AI Systems 

Optical switches are a new kind of switch for AI clusters that handles the switching optically instead of using transceivers to convert photons to electrons, and back again. Optical switching and CPOs work together to allow for more flexibility for reconfigurations, to reduce energy and complexity while also increasing bandwidth. 

There are many competitors within optical switching, with heavyweights Broadcom and Arista coming to mind, yet Lumentum believes their MEMS-Based technology can set them apart. Although the discussion around MEMS can get quite technical, the idea is that Lumentum is a smaller, (potentially) key supplier for customers putting optical circuit switches into their data centers. Another use case for using Lumentum is to rearchitect or write software to enable the optical circuit switching. 

According to a previous earnings call, Lumentum has “already shipped evaluation units to customers who have provided overwhelmingly positive feedback on our performance.” It was also stated that “more meaningful growth will probably be in calendar 2026” for the optical switching circuit products. 

Partner for Nvidia’s Silicon Photonics & Marvell 400G per Lane

Nvidia recently announced 1.6 terabits per second port switches at GTC “to deliver 3.5x energy savings and 10x resilience in AI factories.” Lumentum was named as a partner, among others in the industry.  

This analysis has covered some of Lumentum’s unique advantages and why they would be chosen as a partner, such as: 

  • At the forefront of increasing bandwidths, from data rates of 400G, 800G, and the upcoming 1.6T 
  • At the forefront of faster data pipelines, from 100G per lane (today), 200G per lane (ramping now and into H2 2025), and 400G per lane (into 2026 and beyond) 
  • Indium Phosphide lasers are more expensive due to volumes at the component level compared to silicon — yet is made up for in data centers as InP reduces power consumption. InP also offers higher signal integrity. 
  • 1310nm DFB lasers are high power lasers that enable the transition to co-packaged optics.

Last month, Lumentum announced an InP DFB-MZI optical transmitter in partnership with Marvell to demonstrate 400G-per-lane PAM4. The long acronym refers to a laser-modulator combo — specifically, a DFB laser integrated with a compact Mach-Zehnder (MZI) modulator. This was paired with Marvell’s 400G PAM4 DSP, which operates at 225 Gbaud and enables the high-speed signaling needed to reach 400G per lane. 

The result is a high-performance, power-efficient optical transmitter that outperforms traditional silicon photonics — “particularly in applications where signal integrity and efficiency are critical,” per the press release. The InP DFB-MZI platform sets the stage for architectures expanding to 1.6T and 3.2T bandwidth, which is in the near future, up from the 800G modules in production today. 

Lumentum is Supply Constrained; Building Capacity 

Lumentum’s demand far exceeds supply, including indium phosphide supply and components such as CW lasers. 

The company is expanding its transceiver manufacturing capacity through the construction of a three-story facility and cleanroom in Thailand, which complements existing production lines. The first floor is completed and ready for tool installation.  

As stated, Lumentum began shipping its 200G lane speed externally-modulated lasers (EMLs) to multiple customers in Q2 F2025. The Company aims to increase its EML capacity by 40% YoY from June 2024 to June 2025, followed by another 40% increase by the end of 2025 as its Thailand production accelerates.   

In Q2, the Company invested $65 million in capex to expand cleanroom capacity and increase equipment capacity for InP wafer production to support EML chip manufacturing at its Thailand site. Lumentum has also been invested in expanding its indium phosphide (InP) wafer fabrication facilities, which are critical for producing high-speed lasers like the EMLs used in transceivers.  

CEO Lowe reiterated this in the Q2 F2025 conference call: 

“We're still on track to what we've been saying, which was 40% or higher growth from the June quarter of calendar '24 to the June quarter of calendar '25 and then another 40% by the end of calendar '25. So, that's for all 100 gig and 200 gig. I think we were in our prepared remarks, we talked about being overall supply constrained, not 200 gig because we can start a 200 gig wafer or a 100 gig wafer just the same. So, we're just overall the demand is outstripping even that 40% compounded twice in an 18-month period of time… Our wafer fab expansion plans to enable higher volumes of EMLs and other indium phosphide lasers and photodetectors continues to be on track. We still anticipate that demand for our EML chips will continue to exceed supply, at least into calendar year 2026. We are experiencing strengthening demand for our DCI products as well as long haul transmission and transport solutions.” 

A Note on China Exposure 

Interestingly, Lumentum does not have as much exposure to China whereas most supply chain troubles right now are due to sourcing in this problematic geo-political zone  

This was stated on the last call: 

“We have some production that happens in China and then those components are integrated into bigger components, bigger products at our Thailand facility. And then most of the shipments happen from Thailand, even if they are to the US or to Mexico as well. And then much of the growth that we're seeing is shipments that are coming from Japan as well as from Caswell for the transmission products that we have. So because of that, at least in the short to mid-term, unless policy changes, happen at a government level, we're not expecting much of an impact.” 

This was also stated in the introduction: 

“Second, we are scaling capacity for our highly differentiated laser transmitter chips in our indium phosphide wafer fabs and optical circuit switch and transceiver production capacity in our proven factories outside of China to meet the rising demand.” 

Quite a lot has changed in terms of policy at the government level since the last earnings call. The overall commentary regarding China reliance being minimal on the supply chain side may be true — but tariffs could upend enough of the supply chain to affect Lumentum’s customers (and overall demand). 

Also, despite not relying on manufacturing in China, Lumentum reported Hong Kong as a major customer in the 10-Q at 19% of revenue for LITE, second to United States. Geo-political tensions should cause China/HK to source domestically either voluntarily or through blacklists.  

Source: Lumentum’s 10-QLumentum’s 10-Q 

Telecom Problematic End Market yet DCIs are a Bright Spot 

There has been a steep inventory correction in telecom that has led to substantial revenue decline and significant margin erosion, presenting a major fundamental headwind for data center growth to overcome. However, data center interconnects (DCIs) are helping to drive the turnaround in this otherwise problematic end market.  

Lumentum offers tunable laser and coherent pluggable transceivers for data center interconnects (DCIs). These long-distance data transmissions are traditionally used for telecom purposes and can range up to hundreds of kilometers yet are now seeing demand for data center buildouts.  

In April 2025, Lumentum announced the sampling of new 400/800G ZR+ L-band pluggable transceivers and the general availability of its 800G ZR+ C-band module. The L-band modules effectively double the usable wavelength range and available fiber capacity. By expanding into both C-band and L-band spectrums, Lumentum’s transceivers enable increased fiber capacity, and are able to serve AI and cloud-based applications. These pluggable modules reduce overall system complexity and cost.   

Regarding DCIs, per the previous earnings call: “we're seeing dramatic strength in anything ZR, anything to connect data centers as data centers are being built out, and that can take the form of ZR modules. But given our share of tunable lasers that go into ZRs, that's where we're going to see a dramatic pickup in the telecom side.” 

The most recent update from management is that demand remains very, very strong: “And so the data center interconnect and the networks that are interconnecting these data centers is showing very, very strong demand. And so DCIs and the components as well as ROADMs amplifiers and longer haul coherent transmission connecting further apart data centers is very strong.” 

Revenue Growth Poised to Accelerate  

Q2 FY25 revenue grew by 9.65% YoY and 19.37% QoQ to $402.2 million, driven by strength in its Cloud and Networking segment, beating estimates by 2.87%. 

  • Management guided Q3 revenue between $410 million to $425 million, with a midpoint of $417.5 million for 13.9% growth. 
  • Analysts expect revenue to grow 46.6% in the June quarter to $452 million and 43% YoY in the September quarter to $481.9 million in Q1 F2026. 
  • Management also reaffirmed its commitment to reaching quarterly revenue of $500 million by the end of calendar year 2025, driven by improving trends with its networking equipment manufacturing customers. 

Segments: 

Cloud and Networking Drives Revenue and Margins

In the most recent quarter, Cloud and Networking grew 18.3% YoY and 20.2% QoQ to $339.2 million with management stating they saw “sequential increases in nearly all of our Cloud and Networking product lines.” Cloud and Networking segment profit grew 16.2% for an increase of 330 basis points sequentially and an increase of 610 basis points year-on-year on higher revenue. This segment includes optical transceivers, the datacom chips/lasers that go into optical modules (EMLs, VSCELs, CW lasers, etc) and telecom/data center interconnects. 

Previously, in the September quarter, the segment saw a meaningful inflection with growth of 23% YoY and 11% QoQ. Next quarter, management stated they expect the segment to increase by $25 million QoQ, which would represent growth of 7.3% QoQ – not quite as high as the previous quarter yet it’s been clearly noted in our previous analysis that H2 would be the bigger ramp. Our analysis on delayed Nvidia suppliers also connects some dots on this particular timing for Lumentum. 

According to management commentary, the company shipped record EML units with datacom transceivers shipping to their largest hyperscaler customer and volume production shipments to a new customer.  

The company stated they expect to increase their market share: 

 “Based on the breadth of our 200G EML design wins, we expect to gain additional laser transmitter market share in the upcoming wave of 800G and 1.6T transceivers, utilizing the more efficient 200G EMLs for AI applications. Complementing our EMLs are our new 200G lens integrated photodetector arrays, which adds to our content opportunity in next generation 800G and 1.6T transceivers as well as strengthens our vertical integration strategy for our own cloud modules.” 

Industrial Tech Continues to Contract 

Q2 F2024 Industrial Tech revenue fell (21.4%) YoY and grew 15.4% QoQ to $63 million. The sequential increase was driven by higher industrial laser shipments, partially offset by lower 3D sensing shipments. 

Next quarter, management expects revenues to decline sequentially by $10 million, driven by declines in both commercial lasers and 3D sensing. 

Adj. Margins Bottomed Out and Accelerating  

GAAP Gross margin was 24.7% with adjusted gross margin at 32.3%. Per management, company gross margins will “sequentially increase as manufacturing utilization improves as well as an increase in Datacom laser shipments.” 

  • GAAP operating margin was (12.8%) for operating losses of ($51.6) million. 
  • Adjusted operating margins rose to 7.9% compared to 3% last quarter. It’s expected to further expand to 10% at the midpoint this quarter. 
  • GAAP net margin was (15.1%) for net losses of $61 million 
  • GAAP adjusted net margin was 7.46% for adjusted profits of $30 million 

Adj. EPS Returns to Growth and Accelerates Quickly 

Q2 FY25 adj. EPS improved to $0.42, beating consensus estimates by 16.97%. Management guided Q3 F2025 adj. EPS between $0.47 to $0.53, with a midpoint of $0.53. There is outsized growth in EPS from the rebound on small numbers: 

  • Analysts expect Q4 adj. EPS to grow 957.38% to $0.63  
  • Q1 is expected to see 332.24% to $0.78 

The difference between GAAP and non-GAAP operating margin is due to the stock-based compensation expenses and amortization of acquired intangibles.  

Elevated Capex Spend in Thailand Manufacturing Site Pressuring Free Cash Flow 

Q2 FY25 operating cash flow was $24.3 million or 6.1% of revenue. The Company spent $74 million capex in Q1 FY25 and $64 million in Q2, resulting in sequential negative free cash flow.  

Free cash flow in Q2 was ($15.9 million) or -4% of revenue.   The Company closed the quarter with $896.7 million in cash and cash equivalents and $2.47 billion in debt. This puts the debt-to-equity ratio at 2.75 due to high capex spending — which is high and not ideal in this environment. However, if Lumentum can prove the capex will quickly be converted to revenue, it may become a non-issue by this time next year.  

Lumentum spent $64 million in capex for expanding cleanroom capacity at the Thailand manufacturing site and increasing equipment capacity for indium phosphate wafer production to support EML chip manufacturing. 

Valuation  

Lumentum trades at a forward price-earnings (P/E) of 31.3 and a current PE ratio of 143 (although this looks drastic given the weak bottom line the company is rebounding from). 

The price/sales (P/S) ratio is 2.6 and forward P/S is 2.3. The five-year average P/S ratio is 3.2 

These valuations do not reflect an AI story should Lumentum catch the AI bid (in a bull market) the valuation could be 5-6. This may not be in the near-term given tariff related concerns, weak semiconductor sector performance and Nasdaq entering a bear market officially following the (20%) decline. However, looking into the second half of the year, should conditions improve, Lumentum is capable of trading higher. 

Q&A from Earnings Call 

Yield/Supply Issues are Limiting Growth 

According to the earnings call, the Cloud and Networking segment is expected to increase by $25 million yet the company has “demand that far surpasses that.” Management went on to explain: “we could have probably seen a double-digit increase sequentially quarter-over-quarter had we not had some of those supply chain shortages.” 

It was later more specifically called out as yield issues on the transceiver side, with this comment being in context of the lower margin: “So, yes, so during the quarter, we had some yield issues related to new product ramps within our Transceiver business. That probably was a headwind of anywhere from 100 to 150 basis points.” 

This isn’t exactly surprising given the clear commentary around capex and the need to increase capacity. It was also later stated the supply issues should ease by the June quarter:  

“To your second question, we're gating our module customers’ ability to grow their Transceiver business by the lack of worldwide indium phosphide for EMLs and CW lasers, quite frankly, we're having challenges actually getting enough CW lasers for our own transceivers. So that is actually impacting us on the short-term this quarter, hope to have that resolved in the June quarter. “ 

With even further questioning, it was revealed that hermetic packaging is also creating supply shortages, which refers to sealed enclosures that protect optical components. 

“But the worldwide shortage of things like hermetic packages is creating a challenge for the kinds of volumes that our customers are looking for especially in coherent components and narrow line with lasers. And so, if we have those packages and we could get them more readily, we could grow, as Wajid said, double-digits quarter-on-quarter. It is going to hamper us in both the March quarter as well as the June quarter, and we're working diligently to minimize that impact in the June quarter, but the March quarter is what it is because we need those deliveries now in order to turn products for the quarter and we have a limited supply and ability to get that for example.” 

Hyperscaler Customers 

In the most recent quarter, the 10-Q states that three customers accounted for 16%, 14% and 11% of total revenue with two customers accounting for gross accounts receivable.  

  • In the prior quarter, two customers accounted for 15% and 12% of revenue, respectively.  
  • When comparing to the quarter a year ago, there were three customers at 19%, 13% and 11% of total revenue.  

Management provided the following color in regard to how the ramp is going with the three customers: 

“In Q2, Datacom transceiver revenue grew sequentially as expected, driven by an increase in shipments to our largest cloud hyperscale customer and the start of volume production shipments to one of our new customers we highlighted on prior calls. 

We continue qualification work with the other new customers and expect to start initial volume production during the fourth quarter continuing to ramp through the first half of fiscal '26. Transceiver manufacturing capacity expansion is also progressing as planned.” 

During the call, there was a Q&A exchange that drilled deeper into a potential fourth customer: 

“Ananda Baruah:  

 “On the new ramping customer, can you give us some sense of time frame when you think that it hits run rate and or any context around what run rate I know this is probably kind of somewhat project based, but when you think it hits run rate, what time frame? And then the second one is sounds like this year, calendar year '25, you're not going to run into sort of congestion between your EML chips and your own transceivers since they're largely SiPho (silicon photonics) based right now.”  

Lowe:  

 “As far as the new customer reaching run rate, I'd say that's going to take some time. So don't be raising your projections for that customer. But as I said earlier, we're qualifying a second product there that should come on by the end of the calendar year. So, I'd say around the end of the calendar year, we should be in full motion with that customer. And then as we talked about in the script, the third customer start production in fiscal Q4 and really hit run rate, I'd say, by the fall time. So, September, October. That product is scheduled to ramp significantly faster. So that's the ramp color.”  

Conclusion: 

Lumentum is a small-cap company with differentiated technology that has caught the attention of heavyweight NVIDIA. We’ve been closely monitoring Lumentum for roughly a year, waiting patiently for their EML lasers for 200G to ship, enabling 800G and 1.6T bandwidths. Any progress here should continue into 2026-2027 for 400G data lanes and 3.2T bandwidths. 

Their indium phosphide (InP) laser technology offers significant power efficiency advantages over traditional silicon photonics. This is increasingly important as power consumption becomes a central concern in scaling AI data centers. Lumentum’s collaboration with NVIDIA, integrating their high-efficiency lasers into NVIDIA’s Spectrum-X and Quantum-X photonics networking switches, is a nod towards the EML lasers and their ultra high power lasers for CPOs playing a critical role in future architectures. 

With that said, Lumentum is in the high risk bucket due to being a small cap. The stock requires speculation as to when a shift in fundamentals will occur. To date, we have seen an inflection for one quarter, yet we need more evidence before an inflection becomes a meaningful trend. The sharp acceleration provided for in analyst consensus in H2 could wane if supply chain troubles trickle down and result in slower sourcing for AI systems.  

Ultimately, it’s well worth our time to earmark companies like Lumentum and identify potential entry targets.  

The I/O Fund recently launched our new Discovery tier which surfaces new ideas the I/O Fund does not own yet at a pace of 30-40 new stocks per year. Coverage will include AI hardware, AI software, crypto and more, from a leading tech portfolio.  

Sample research we published in March and April:   

  • Key supplier to TSMC’s new high-growth platform called Compact Universal Photonic Engine (COUPE)  
  • Nuclear and natural gas supplier for AI data centers   
  • A breakdown of the risks and opportunities for the biggest IPO in the AI sector

Please note: The I/O Fund conducts research and draws conclusions for the Fund’s positions. We then share that information with our readers. This is not a guarantee of a stock’s performance. Please consult your personal financial advisor before buying any stock in the companies mentioned in this analysis.

Recommended Reading:

Lumentum at Inflection Point with 20% QoQ Growth in AI-Related Segment

Posted on by io-fund

Lumentum has been on our radar for some time as the company supplies components for datacom transceivers and optical interconnects. Lumentum is a small-cap company with differentiated technology that has caught the attention of heavyweight NVIDIA. We’ve been closely monitoring Lumentum for roughly a year, waiting patiently for their EML lasers for 200G to ship, enabling 800G and 1.6T bandwidths. Any progress here should continue into 2026-2027 for 400G data lanes and 3.2T bandwidths. 

Optical interconnects help data centers accelerate data throughput between data centers and inside the data center between servers or racks, while reducing latency and power consumption. AI is driving cloud demand higher from the hyperscalers, leading to more data being created and processed, thus helping drive a need for these interconnects to meet demand for high-speed, low power data transmission in data centers. 

The company supplies three types of lasers for datacom transceivers: VSCELs, CW lasers for silicon photonics and EML-based lasers. Per our previous analysis, the 200G EMLs are what is expected to drive an inflection in 2025.  

In the most recent quarter, Lumentum reported their Cloud and Networking segment grew 18.3% YoY and 20.2% QoQ to $339.2 million. This caused total revenue to grow 9.7% yet total revenue is expected to accelerate to 46.6% and 43% in the second half of the year.  

More on Lumentum’s EML Lasers: 

EMLs were traditionally used by telecom customers, yet became attractive for AI servers due to meeting the 200G per second speeds necessary for 1.6T optical modules to support AI models. These are called single mode optics, made of Indium Phosphide, which has been used instead of silicon for long-haul networking due to being a superior choice for optical functions, such as enabling the laser, modulator, photodetector and amplifier.  

InP is more expensive at the component level as four EMLs are needed compared to two lower-cost CW lasers for silicon photonics modules, yet this difference at the component level can be made up for in data centers as InP reduces power consumption. 

In the December quarter, Lumentum stated they began shipping its 200G lane speed externally-modulated lasers (EMLs) to multiple customers. The Company stated they would increase its EML capacity by 40% YoY from June 2024 to June 2025, followed by another 40% increase by the end of calendar 2025 as its Thailand production accelerates – representing compounding growth in an 18-month time frame:   

“So, we're just overall the demand is outstripping even that 40% compounded twice in an 18 month period of time. We do have additional capacity expansion beyond the end of calendar '25 obviously and those are investments that we've made over the past several quarters that come online in calendar '25, which will give us increased capacity as well.” 

Lumentum’s indium phosphide 100G EMLs (Externally-Modulated Lasers) are being shipped and used in leading single-mode 400G and 800G optical transceivers. These customers are now designing the company’s 200G EMLs into their next generation of transceivers. Already being a lead supplier for 100G EML transceivers, the company is set up to be in pole position for the 200G EML transceivers. 

Per the most recent earnings call: “We achieved another record for EML unit shipments in Q2 and began delivering 200G lane speed EMLs to multiple customers.  

Based on the breadth of our 200G EML design wins, we expect to gain additional laser transmitter market share in the upcoming wave of 800G and 1.6T transceivers, utilizing the more efficient 200G EMLs for AI applications. Complementing our EMLs are our new 200G lens integrated photodetector arrays, which adds to our content opportunity in next generation 800G and 1.6T transceivers as well as strengthens our vertical integration strategy for our own cloud modules.” 

Specifically, the company’s experience in InP long haul transceivers is being tapped as AI servers scale, especially since InP reduces power consumption compared to silicon. Lumentum is also working on higher speed optical links, including 400G per lane. This means more data, fewer lanes if you compare it to 200G per lane. This is important especially as data centers will eventually move not only to 1.6 terabits per second (1.6T) (4X400G or 8X200G) but also 3.2T (8X400G).  

Management also stated the new 200G lens integrated photodetector arrays “adds to our content opportunity in next generation 800G and 1.6T transceivers.” 

Lumentum’s Co-Packaged Optics Opportunity for High Power Lasers:  

Silicon photonics are the only viable choice for rack-to-rack and across the data center due to the need for high bandwidth and lower power at high speeds. There is also low-loss over long distances with optical fiber, which refers to preserving the original signal, whereas copper sees signal degradation over longer distances.  

Where there is a debate and an important shift occurring is in the networking between chips and inside the rack. We’ve recently covered the benefits of co-packaged optics as a replacement for pluggable transceivers when networking between chips. At 800G and 1.6T speeds, the electrical signals over power circuit boards (PCBs) run too hot, are power hungry and create loss for the signal. The overall goal is to move the optics closer in proximity to overcome scaling issues when increasing electrical speed. 

Nvidia stated that by replacing pluggable optics with silicon photonics on the package, it can “deliver 3.5x more power efficiency, 63x greater signal integrity, 10x better network resiliency at scale and 1.3x faster deployment compared with traditional methods.”  

Nvidia VP Ian Buck stated at GTC that the CPO switches help reduce power consumption by eliminating the need for external lasers and pluggable transceivers to achieve a significant reduction in power from 39 watts to 9 watts. Buck explained that this “gives you that benefit from going from 39 watts of power down to only 9 watts of power for the same number of ports, and that's huge. It doesn't sound like 39 sounds a lot. But if you get 400,000 GPUs in an AI supercomputer, there's like 24 megawatts of lasers like so that's a lot of laser light that could be optimized and made more efficient.” 

While Lumentum may see some decline in its pluggable optics transceiver business as the industry shifts, it is also well-positioned to benefit from the transition. The company is a key supplier of high-power lasers required for co-packaged optics (CPO), which could offset near-term losses. Taken together, the impact should be at the onset net neutral — with (potentially significant) upside over time from growing demand for EML lasers, the transition to 200G and 400G per lane, plus optical switching (see below). 

Regarding CPOs, here is what was stated on the most recent call: “But I'd also highlight for Lumentum a little bit of a unique situation where we're more modest share on transceivers today and share gaining, if you will, over time. So even if there is some cannibalization in the mid to long-term, we don't think it impacts our transceiver opportunity and creates an expanded opportunity for high power lasers.” 

Nearly a year ago, Lumentum announced an ultra-high output power 1310 nm DFB laser in beta, stating it was designed to reduce the “required number of lasers” while boosting efficiency and reliability in large scale AI/ML infrastructure. 

It was also shared that Lumentum is shipping preproduction volumes now related to CPOs: 

 ‘As part of one collaboration, we began shipping preproduction volumes of our unique ultra-high power lasers to an AI infrastructure customer for a proprietary interconnect solution in Q2 and have received follow-on orders as well as excellent feedback on the product's performance. This is a very exciting opportunity.” 

It was later teased out on the earnings call that 2026 could be a big year for Lumentum’s high power laser: “I would say preproduction volume with forecasts for more meaningful volume throughout calendar '25 and then real extremely high volume in calendar '26. So, feedback is very positive. The performance is very unique and the line width that we produce with the laser is superior to anything in the market.”

CPO (with Switch ASIC) + Optical Switch + GPU Clusters = The Future of Scale Up AI Systems 

Optical switches are a new kind of switch for AI clusters that handles the switching optically instead of using transceivers to convert photons to electrons, and back again. Optical switching and CPOs work together to allow for more flexibility for reconfigurations, to reduce energy and complexity while also increasing bandwidth. 

There are many competitors within optical switching, with heavyweights Broadcom and Arista coming to mind, yet Lumentum believes their MEMS-Based technology can set them apart. Although the discussion around MEMS can get quite technical, the idea is that Lumentum is a smaller, (potentially) key supplier for customers putting optical circuit switches into their data centers. Another use case for using Lumentum is to rearchitect or write software to enable the optical circuit switching. 

According to a previous earnings call, Lumentum has “already shipped evaluation units to customers who have provided overwhelmingly positive feedback on our performance.” It was also stated that “more meaningful growth will probably be in calendar 2026” for the optical switching circuit products. 

Partner for Nvidia’s Silicon Photonics & Marvell 400G per Lane

Nvidia recently announced 1.6 terabits per second port switches at GTC “to deliver 3.5x energy savings and 10x resilience in AI factories.” Lumentum was named as a partner, among others in the industry.  

This analysis has covered some of Lumentum’s unique advantages and why they would be chosen as a partner, such as: 

  • At the forefront of increasing bandwidths, from data rates of 400G, 800G, and the upcoming 1.6T 
  • At the forefront of faster data pipelines, from 100G per lane (today), 200G per lane (ramping now and into H2 2025), and 400G per lane (into 2026 and beyond) 
  • Indium Phosphide lasers are more expensive due to volumes at the component level compared to silicon — yet is made up for in data centers as InP reduces power consumption. InP also offers higher signal integrity. 
  • 1310nm DFB lasers are high power lasers that enable the transition to co-packaged optics.

Last month, Lumentum announced an InP DFB-MZI optical transmitter in partnership with Marvell to demonstrate 400G-per-lane PAM4. The long acronym refers to a laser-modulator combo — specifically, a DFB laser integrated with a compact Mach-Zehnder (MZI) modulator. This was paired with Marvell’s 400G PAM4 DSP, which operates at 225 Gbaud and enables the high-speed signaling needed to reach 400G per lane. 

The result is a high-performance, power-efficient optical transmitter that outperforms traditional silicon photonics — “particularly in applications where signal integrity and efficiency are critical,” per the press release. The InP DFB-MZI platform sets the stage for architectures expanding to 1.6T and 3.2T bandwidth, which is in the near future, up from the 800G modules in production today. 

Lumentum is Supply Constrained; Building Capacity 

Lumentum’s demand far exceeds supply, including indium phosphide supply and components such as CW lasers. 

The company is expanding its transceiver manufacturing capacity through the construction of a three-story facility and cleanroom in Thailand, which complements existing production lines. The first floor is completed and ready for tool installation.  

As stated, Lumentum began shipping its 200G lane speed externally-modulated lasers (EMLs) to multiple customers in Q2 F2025. The Company aims to increase its EML capacity by 40% YoY from June 2024 to June 2025, followed by another 40% increase by the end of 2025 as its Thailand production accelerates.   

In Q2, the Company invested $65 million in capex to expand cleanroom capacity and increase equipment capacity for InP wafer production to support EML chip manufacturing at its Thailand site. Lumentum has also been invested in expanding its indium phosphide (InP) wafer fabrication facilities, which are critical for producing high-speed lasers like the EMLs used in transceivers.  

CEO Lowe reiterated this in the Q2 F2025 conference call: 

“We're still on track to what we've been saying, which was 40% or higher growth from the June quarter of calendar '24 to the June quarter of calendar '25 and then another 40% by the end of calendar '25. So, that's for all 100 gig and 200 gig. I think we were in our prepared remarks, we talked about being overall supply constrained, not 200 gig because we can start a 200 gig wafer or a 100 gig wafer just the same. So, we're just overall the demand is outstripping even that 40% compounded twice in an 18-month period of time… Our wafer fab expansion plans to enable higher volumes of EMLs and other indium phosphide lasers and photodetectors continues to be on track. We still anticipate that demand for our EML chips will continue to exceed supply, at least into calendar year 2026. We are experiencing strengthening demand for our DCI products as well as long haul transmission and transport solutions.” 

A Note on China Exposure 

Interestingly, Lumentum does not have as much exposure to China whereas most supply chain troubles right now are due to sourcing in this problematic geo-political zone  

This was stated on the last call: 

“We have some production that happens in China and then those components are integrated into bigger components, bigger products at our Thailand facility. And then most of the shipments happen from Thailand, even if they are to the US or to Mexico as well. And then much of the growth that we're seeing is shipments that are coming from Japan as well as from Caswell for the transmission products that we have. So because of that, at least in the short to mid-term, unless policy changes, happen at a government level, we're not expecting much of an impact.” 

This was also stated in the introduction: 

“Second, we are scaling capacity for our highly differentiated laser transmitter chips in our indium phosphide wafer fabs and optical circuit switch and transceiver production capacity in our proven factories outside of China to meet the rising demand.” 

Quite a lot has changed in terms of policy at the government level since the last earnings call. The overall commentary regarding China reliance being minimal on the supply chain side may be true — but tariffs could upend enough of the supply chain to affect Lumentum’s customers (and overall demand). 

Also, despite not relying on manufacturing in China, Lumentum reported Hong Kong as a major customer in the 10-Q at 19% of revenue for LITE, second to United States. Geo-political tensions should cause China/HK to source domestically either voluntarily or through blacklists.  

Source: Lumentum’s 10-QLumentum’s 10-Q 

Telecom Problematic End Market yet DCIs are a Bright Spot 

There has been a steep inventory correction in telecom that has led to substantial revenue decline and significant margin erosion, presenting a major fundamental headwind for data center growth to overcome. However, data center interconnects (DCIs) are helping to drive the turnaround in this otherwise problematic end market.  

Lumentum offers tunable laser and coherent pluggable transceivers for data center interconnects (DCIs). These long-distance data transmissions are traditionally used for telecom purposes and can range up to hundreds of kilometers yet are now seeing demand for data center buildouts.  

In April 2025, Lumentum announced the sampling of new 400/800G ZR+ L-band pluggable transceivers and the general availability of its 800G ZR+ C-band module. The L-band modules effectively double the usable wavelength range and available fiber capacity. By expanding into both C-band and L-band spectrums, Lumentum’s transceivers enable increased fiber capacity, and are able to serve AI and cloud-based applications. These pluggable modules reduce overall system complexity and cost.   

Regarding DCIs, per the previous earnings call: “we're seeing dramatic strength in anything ZR, anything to connect data centers as data centers are being built out, and that can take the form of ZR modules. But given our share of tunable lasers that go into ZRs, that's where we're going to see a dramatic pickup in the telecom side.” 

The most recent update from management is that demand remains very, very strong: “And so the data center interconnect and the networks that are interconnecting these data centers is showing very, very strong demand. And so DCIs and the components as well as ROADMs amplifiers and longer haul coherent transmission connecting further apart data centers is very strong.” 

Revenue Growth Poised to Accelerate  

Q2 FY25 revenue grew by 9.65% YoY and 19.37% QoQ to $402.2 million, driven by strength in its Cloud and Networking segment, beating estimates by 2.87%. 

  • Management guided Q3 revenue between $410 million to $425 million, with a midpoint of $417.5 million for 13.9% growth. 
  • Analysts expect revenue to grow 46.6% in the June quarter to $452 million and 43% YoY in the September quarter to $481.9 million in Q1 F2026. 
  • Management also reaffirmed its commitment to reaching quarterly revenue of $500 million by the end of calendar year 2025, driven by improving trends with its networking equipment manufacturing customers. 

Segments: 

Cloud and Networking Drives Revenue and Margins

In the most recent quarter, Cloud and Networking grew 18.3% YoY and 20.2% QoQ to $339.2 million with management stating they saw “sequential increases in nearly all of our Cloud and Networking product lines.” Cloud and Networking segment profit grew 16.2% for an increase of 330 basis points sequentially and an increase of 610 basis points year-on-year on higher revenue. This segment includes optical transceivers, the datacom chips/lasers that go into optical modules (EMLs, VSCELs, CW lasers, etc) and telecom/data center interconnects. 

Previously, in the September quarter, the segment saw a meaningful inflection with growth of 23% YoY and 11% QoQ. Next quarter, management stated they expect the segment to increase by $25 million QoQ, which would represent growth of 7.3% QoQ – not quite as high as the previous quarter yet it’s been clearly noted in our previous analysis that H2 would be the bigger ramp. Our analysis on delayed Nvidia suppliers also connects some dots on this particular timing for Lumentum. 

According to management commentary, the company shipped record EML units with datacom transceivers shipping to their largest hyperscaler customer and volume production shipments to a new customer.  

The company stated they expect to increase their market share: 

 “Based on the breadth of our 200G EML design wins, we expect to gain additional laser transmitter market share in the upcoming wave of 800G and 1.6T transceivers, utilizing the more efficient 200G EMLs for AI applications. Complementing our EMLs are our new 200G lens integrated photodetector arrays, which adds to our content opportunity in next generation 800G and 1.6T transceivers as well as strengthens our vertical integration strategy for our own cloud modules.” 

Industrial Tech Continues to Contract 

Q2 F2024 Industrial Tech revenue fell (21.4%) YoY and grew 15.4% QoQ to $63 million. The sequential increase was driven by higher industrial laser shipments, partially offset by lower 3D sensing shipments. 

Next quarter, management expects revenues to decline sequentially by $10 million, driven by declines in both commercial lasers and 3D sensing. 

Adj. Margins Bottomed Out and Accelerating  

GAAP Gross margin was 24.7% with adjusted gross margin at 32.3%. Per management, company gross margins will “sequentially increase as manufacturing utilization improves as well as an increase in Datacom laser shipments.” 

  • GAAP operating margin was (12.8%) for operating losses of ($51.6) million. 
  • Adjusted operating margins rose to 7.9% compared to 3% last quarter. It’s expected to further expand to 10% at the midpoint this quarter. 
  • GAAP net margin was (15.1%) for net losses of $61 million 
  • GAAP adjusted net margin was 7.46% for adjusted profits of $30 million 

Adj. EPS Returns to Growth and Accelerates Quickly 

Q2 FY25 adj. EPS improved to $0.42, beating consensus estimates by 16.97%. Management guided Q3 F2025 adj. EPS between $0.47 to $0.53, with a midpoint of $0.53. There is outsized growth in EPS from the rebound on small numbers: 

  • Analysts expect Q4 adj. EPS to grow 957.38% to $0.63  
  • Q1 is expected to see 332.24% to $0.78 

The difference between GAAP and non-GAAP operating margin is due to the stock-based compensation expenses and amortization of acquired intangibles.  

Elevated Capex Spend in Thailand Manufacturing Site Pressuring Free Cash Flow 

Q2 FY25 operating cash flow was $24.3 million or 6.1% of revenue. The Company spent $74 million capex in Q1 FY25 and $64 million in Q2, resulting in sequential negative free cash flow.  

Free cash flow in Q2 was ($15.9 million) or -4% of revenue.   The Company closed the quarter with $896.7 million in cash and cash equivalents and $2.47 billion in debt. This puts the debt-to-equity ratio at 2.75 due to high capex spending — which is high and not ideal in this environment. However, if Lumentum can prove the capex will quickly be converted to revenue, it may become a non-issue by this time next year.  

Lumentum spent $64 million in capex for expanding cleanroom capacity at the Thailand manufacturing site and increasing equipment capacity for indium phosphate wafer production to support EML chip manufacturing. 

Valuation  

Lumentum trades at a forward price-earnings (P/E) of 31.3 and a current PE ratio of 143 (although this looks drastic given the weak bottom line the company is rebounding from). 

The price/sales (P/S) ratio is 2.6 and forward P/S is 2.3. The five-year average P/S ratio is 3.2 

These valuations do not reflect an AI story should Lumentum catch the AI bid (in a bull market) the valuation could be 5-6. This may not be in the near-term given tariff related concerns, weak semiconductor sector performance and Nasdaq entering a bear market officially following the (20%) decline. However, looking into the second half of the year, should conditions improve, Lumentum is capable of trading higher. 

Q&A from Earnings Call 

Yield/Supply Issues are Limiting Growth 

According to the earnings call, the Cloud and Networking segment is expected to increase by $25 million yet the company has “demand that far surpasses that.” Management went on to explain: “we could have probably seen a double-digit increase sequentially quarter-over-quarter had we not had some of those supply chain shortages.” 

It was later more specifically called out as yield issues on the transceiver side, with this comment being in context of the lower margin: “So, yes, so during the quarter, we had some yield issues related to new product ramps within our Transceiver business. That probably was a headwind of anywhere from 100 to 150 basis points.” 

This isn’t exactly surprising given the clear commentary around capex and the need to increase capacity. It was also later stated the supply issues should ease by the June quarter:  

“To your second question, we're gating our module customers’ ability to grow their Transceiver business by the lack of worldwide indium phosphide for EMLs and CW lasers, quite frankly, we're having challenges actually getting enough CW lasers for our own transceivers. So that is actually impacting us on the short-term this quarter, hope to have that resolved in the June quarter. “ 

With even further questioning, it was revealed that hermetic packaging is also creating supply shortages, which refers to sealed enclosures that protect optical components. 

“But the worldwide shortage of things like hermetic packages is creating a challenge for the kinds of volumes that our customers are looking for especially in coherent components and narrow line with lasers. And so, if we have those packages and we could get them more readily, we could grow, as Wajid said, double-digits quarter-on-quarter. It is going to hamper us in both the March quarter as well as the June quarter, and we're working diligently to minimize that impact in the June quarter, but the March quarter is what it is because we need those deliveries now in order to turn products for the quarter and we have a limited supply and ability to get that for example.” 

Hyperscaler Customers 

In the most recent quarter, the 10-Q states that three customers accounted for 16%, 14% and 11% of total revenue with two customers accounting for gross accounts receivable.  

  • In the prior quarter, two customers accounted for 15% and 12% of revenue, respectively.  
  • When comparing to the quarter a year ago, there were three customers at 19%, 13% and 11% of total revenue.  

Management provided the following color in regard to how the ramp is going with the three customers: 

“In Q2, Datacom transceiver revenue grew sequentially as expected, driven by an increase in shipments to our largest cloud hyperscale customer and the start of volume production shipments to one of our new customers we highlighted on prior calls. 

We continue qualification work with the other new customers and expect to start initial volume production during the fourth quarter continuing to ramp through the first half of fiscal '26. Transceiver manufacturing capacity expansion is also progressing as planned.” 

During the call, there was a Q&A exchange that drilled deeper into a potential fourth customer: 

“Ananda Baruah:  

 “On the new ramping customer, can you give us some sense of time frame when you think that it hits run rate and or any context around what run rate I know this is probably kind of somewhat project based, but when you think it hits run rate, what time frame? And then the second one is sounds like this year, calendar year '25, you're not going to run into sort of congestion between your EML chips and your own transceivers since they're largely SiPho (silicon photonics) based right now.”  

Lowe:  

 “As far as the new customer reaching run rate, I'd say that's going to take some time. So don't be raising your projections for that customer. But as I said earlier, we're qualifying a second product there that should come on by the end of the calendar year. So, I'd say around the end of the calendar year, we should be in full motion with that customer. And then as we talked about in the script, the third customer start production in fiscal Q4 and really hit run rate, I'd say, by the fall time. So, September, October. That product is scheduled to ramp significantly faster. So that's the ramp color.”  

Conclusion: 

Lumentum is a small-cap company with differentiated technology that has caught the attention of heavyweight NVIDIA. We’ve been closely monitoring Lumentum for roughly a year, waiting patiently for their EML lasers for 200G to ship, enabling 800G and 1.6T bandwidths. Any progress here should continue into 2026-2027 for 400G data lanes and 3.2T bandwidths. 

Their indium phosphide (InP) laser technology offers significant power efficiency advantages over traditional silicon photonics. This is increasingly important as power consumption becomes a central concern in scaling AI data centers. Lumentum’s collaboration with NVIDIA, integrating their high-efficiency lasers into NVIDIA’s Spectrum-X and Quantum-X photonics networking switches, is a nod towards the EML lasers and their ultra high power lasers for CPOs playing a critical role in future architectures. 

With that said, Lumentum is in the high risk bucket due to being a small cap. The stock requires speculation as to when a shift in fundamentals will occur. To date, we have seen an inflection for one quarter, yet we need more evidence before an inflection becomes a meaningful trend. The sharp acceleration provided for in analyst consensus in H2 could wane if supply chain troubles trickle down and result in slower sourcing for AI systems.  

Ultimately, it’s well worth our time to earmark companies like Lumentum and identify potential entry targets.  

The I/O Fund recently launched our new Discovery tier which surfaces new ideas the I/O Fund does not own yet at a pace of 30-40 new stocks per year. Coverage will include AI hardware, AI software, crypto and more, from a leading tech portfolio. new Discovery tier which surfaces new ideas the I/O Fund does not own yet at a pace of 30-40 new stocks per year. Coverage will include AI hardware, AI software, crypto and more, from a leading tech portfolio.  

Sample research we published in March and April:   

  • Key supplier to TSMC’s new high-growth platform called Compact Universal Photonic Engine (COUPE)  
  • Nuclear and natural gas supplier for AI data centers   
  • A breakdown of the risks and opportunities for the biggest IPO in the AI sector

Please note: The I/O Fund conducts research and draws conclusions for the Fund’s positions. We then share that information with our readers. This is not a guarantee of a stock’s performance. Please consult your personal financial advisor before buying any stock in the companies mentioned in this analysis.

Recommended Reading:

Lumentum at Inflection Point with 20% QoQ Growth in AI-Related Segment

Posted on by io-fund

Lumentum has been on our radar for some time as the company supplies components for datacom transceivers and optical interconnects. Lumentum is a small-cap company with differentiated technology that has caught the attention of heavyweight NVIDIA. We’ve been closely monitoring Lumentum for roughly a year, waiting patiently for their EML lasers for 200G to ship, enabling 800G and 1.6T bandwidths. Any progress here should continue into 2026-2027 for 400G data lanes and 3.2T bandwidths. 

Optical interconnects help data centers accelerate data throughput between data centers and inside the data center between servers or racks, while reducing latency and power consumption. AI is driving cloud demand higher from the hyperscalers, leading to more data being created and processed, thus helping drive a need for these interconnects to meet demand for high-speed, low power data transmission in data centers. 

The company supplies three types of lasers for datacom transceivers: VSCELs, CW lasers for silicon photonics and EML-based lasers. Per our previous analysis, the 200G EMLs are what is expected to drive an inflection in 2025.  

In the most recent quarter, Lumentum reported their Cloud and Networking segment grew 18.3% YoY and 20.2% QoQ to $339.2 million. This caused total revenue to grow 9.7% yet total revenue is expected to accelerate to 46.6% and 43% in the second half of the year.  

More on Lumentum’s EML Lasers: 

EMLs were traditionally used by telecom customers, yet became attractive for AI servers due to meeting the 200G per second speeds necessary for 1.6T optical modules to support AI models. These are called single mode optics, made of Indium Phosphide, which has been used instead of silicon for long-haul networking due to being a superior choice for optical functions, such as enabling the laser, modulator, photodetector and amplifier.  

InP is more expensive at the component level as four EMLs are needed compared to two lower-cost CW lasers for silicon photonics modules, yet this difference at the component level can be made up for in data centers as InP reduces power consumption. 

In the December quarter, Lumentum stated they began shipping its 200G lane speed externally-modulated lasers (EMLs) to multiple customers. The Company stated they would increase its EML capacity by 40% YoY from June 2024 to June 2025, followed by another 40% increase by the end of calendar 2025 as its Thailand production accelerates – representing compounding growth in an 18-month time frame:   

“So, we're just overall the demand is outstripping even that 40% compounded twice in an 18 month period of time. We do have additional capacity expansion beyond the end of calendar '25 obviously and those are investments that we've made over the past several quarters that come online in calendar '25, which will give us increased capacity as well.” 

Lumentum’s indium phosphide 100G EMLs (Externally-Modulated Lasers) are being shipped and used in leading single-mode 400G and 800G optical transceivers. These customers are now designing the company’s 200G EMLs into their next generation of transceivers. Already being a lead supplier for 100G EML transceivers, the company is set up to be in pole position for the 200G EML transceivers. 

Per the most recent earnings call: “We achieved another record for EML unit shipments in Q2 and began delivering 200G lane speed EMLs to multiple customers.  

Based on the breadth of our 200G EML design wins, we expect to gain additional laser transmitter market share in the upcoming wave of 800G and 1.6T transceivers, utilizing the more efficient 200G EMLs for AI applications. Complementing our EMLs are our new 200G lens integrated photodetector arrays, which adds to our content opportunity in next generation 800G and 1.6T transceivers as well as strengthens our vertical integration strategy for our own cloud modules.” 

Specifically, the company’s experience in InP long haul transceivers is being tapped as AI servers scale, especially since InP reduces power consumption compared to silicon. Lumentum is also working on higher speed optical links, including 400G per lane. This means more data, fewer lanes if you compare it to 200G per lane. This is important especially as data centers will eventually move not only to 1.6 terabits per second (1.6T) (4X400G or 8X200G) but also 3.2T (8X400G).  

Management also stated the new 200G lens integrated photodetector arrays “adds to our content opportunity in next generation 800G and 1.6T transceivers.” 

Lumentum’s Co-Packaged Optics Opportunity for High Power Lasers:  

Silicon photonics are the only viable choice for rack-to-rack and across the data center due to the need for high bandwidth and lower power at high speeds. There is also low-loss over long distances with optical fiber, which refers to preserving the original signal, whereas copper sees signal degradation over longer distances.  

Where there is a debate and an important shift occurring is in the networking between chips and inside the rack. We’ve recently covered the benefits of co-packaged optics as a replacement for pluggable transceivers when networking between chips. At 800G and 1.6T speeds, the electrical signals over power circuit boards (PCBs) run too hot, are power hungry and create loss for the signal. The overall goal is to move the optics closer in proximity to overcome scaling issues when increasing electrical speed. 

Nvidia stated that by replacing pluggable optics with silicon photonics on the package, it can “deliver 3.5x more power efficiency, 63x greater signal integrity, 10x better network resiliency at scale and 1.3x faster deployment compared with traditional methods.”  

Nvidia VP Ian Buck stated at GTC that the CPO switches help reduce power consumption by eliminating the need for external lasers and pluggable transceivers to achieve a significant reduction in power from 39 watts to 9 watts. Buck explained that this “gives you that benefit from going from 39 watts of power down to only 9 watts of power for the same number of ports, and that's huge. It doesn't sound like 39 sounds a lot. But if you get 400,000 GPUs in an AI supercomputer, there's like 24 megawatts of lasers like so that's a lot of laser light that could be optimized and made more efficient.” 

While Lumentum may see some decline in its pluggable optics transceiver business as the industry shifts, it is also well-positioned to benefit from the transition. The company is a key supplier of high-power lasers required for co-packaged optics (CPO), which could offset near-term losses. Taken together, the impact should be at the onset net neutral — with (potentially significant) upside over time from growing demand for EML lasers, the transition to 200G and 400G per lane, plus optical switching (see below). 

Regarding CPOs, here is what was stated on the most recent call: “But I'd also highlight for Lumentum a little bit of a unique situation where we're more modest share on transceivers today and share gaining, if you will, over time. So even if there is some cannibalization in the mid to long-term, we don't think it impacts our transceiver opportunity and creates an expanded opportunity for high power lasers.” 

Nearly a year ago, Lumentum announced an ultra-high output power 1310 nm DFB laser in beta, stating it was designed to reduce the “required number of lasers” while boosting efficiency and reliability in large scale AI/ML infrastructure. 

It was also shared that Lumentum is shipping preproduction volumes now related to CPOs: 

 ‘As part of one collaboration, we began shipping preproduction volumes of our unique ultra-high power lasers to an AI infrastructure customer for a proprietary interconnect solution in Q2 and have received follow-on orders as well as excellent feedback on the product's performance. This is a very exciting opportunity.” 

It was later teased out on the earnings call that 2026 could be a big year for Lumentum’s high power laser: “I would say preproduction volume with forecasts for more meaningful volume throughout calendar '25 and then real extremely high volume in calendar '26. So, feedback is very positive. The performance is very unique and the line width that we produce with the laser is superior to anything in the market.”

CPO (with Switch ASIC) + Optical Switch + GPU Clusters = The Future of Scale Up AI Systems 

Optical switches are a new kind of switch for AI clusters that handles the switching optically instead of using transceivers to convert photons to electrons, and back again. Optical switching and CPOs work together to allow for more flexibility for reconfigurations, to reduce energy and complexity while also increasing bandwidth. 

There are many competitors within optical switching, with heavyweights Broadcom and Arista coming to mind, yet Lumentum believes their MEMS-Based technology can set them apart. Although the discussion around MEMS can get quite technical, the idea is that Lumentum is a smaller, (potentially) key supplier for customers putting optical circuit switches into their data centers. Another use case for using Lumentum is to rearchitect or write software to enable the optical circuit switching. 

According to a previous earnings call, Lumentum has “already shipped evaluation units to customers who have provided overwhelmingly positive feedback on our performance.” It was also stated that “more meaningful growth will probably be in calendar 2026” for the optical switching circuit products. 

Partner for Nvidia’s Silicon Photonics & Marvell 400G per Lane

Nvidia recently announced 1.6 terabits per second port switches at GTC “to deliver 3.5x energy savings and 10x resilience in AI factories.” Lumentum was named as a partner, among others in the industry.  

This analysis has covered some of Lumentum’s unique advantages and why they would be chosen as a partner, such as: 

  • At the forefront of increasing bandwidths, from data rates of 400G, 800G, and the upcoming 1.6T 
  • At the forefront of faster data pipelines, from 100G per lane (today), 200G per lane (ramping now and into H2 2025), and 400G per lane (into 2026 and beyond) 
  • Indium Phosphide lasers are more expensive due to volumes at the component level compared to silicon — yet is made up for in data centers as InP reduces power consumption. InP also offers higher signal integrity. 
  • 1310nm DFB lasers are high power lasers that enable the transition to co-packaged optics.

Last month, Lumentum announced an InP DFB-MZI optical transmitter in partnership with Marvell to demonstrate 400G-per-lane PAM4. The long acronym refers to a laser-modulator combo — specifically, a DFB laser integrated with a compact Mach-Zehnder (MZI) modulator. This was paired with Marvell’s 400G PAM4 DSP, which operates at 225 Gbaud and enables the high-speed signaling needed to reach 400G per lane. 

The result is a high-performance, power-efficient optical transmitter that outperforms traditional silicon photonics — “particularly in applications where signal integrity and efficiency are critical,” per the press release. The InP DFB-MZI platform sets the stage for architectures expanding to 1.6T and 3.2T bandwidth, which is in the near future, up from the 800G modules in production today. 

Lumentum is Supply Constrained; Building Capacity 

Lumentum’s demand far exceeds supply, including indium phosphide supply and components such as CW lasers. 

The company is expanding its transceiver manufacturing capacity through the construction of a three-story facility and cleanroom in Thailand, which complements existing production lines. The first floor is completed and ready for tool installation.  

As stated, Lumentum began shipping its 200G lane speed externally-modulated lasers (EMLs) to multiple customers in Q2 F2025. The Company aims to increase its EML capacity by 40% YoY from June 2024 to June 2025, followed by another 40% increase by the end of 2025 as its Thailand production accelerates.   

In Q2, the Company invested $65 million in capex to expand cleanroom capacity and increase equipment capacity for InP wafer production to support EML chip manufacturing at its Thailand site. Lumentum has also been invested in expanding its indium phosphide (InP) wafer fabrication facilities, which are critical for producing high-speed lasers like the EMLs used in transceivers.  

CEO Lowe reiterated this in the Q2 F2025 conference call: 

“We're still on track to what we've been saying, which was 40% or higher growth from the June quarter of calendar '24 to the June quarter of calendar '25 and then another 40% by the end of calendar '25. So, that's for all 100 gig and 200 gig. I think we were in our prepared remarks, we talked about being overall supply constrained, not 200 gig because we can start a 200 gig wafer or a 100 gig wafer just the same. So, we're just overall the demand is outstripping even that 40% compounded twice in an 18-month period of time… Our wafer fab expansion plans to enable higher volumes of EMLs and other indium phosphide lasers and photodetectors continues to be on track. We still anticipate that demand for our EML chips will continue to exceed supply, at least into calendar year 2026. We are experiencing strengthening demand for our DCI products as well as long haul transmission and transport solutions.” 

A Note on China Exposure 

Interestingly, Lumentum does not have as much exposure to China whereas most supply chain troubles right now are due to sourcing in this problematic geo-political zone  

This was stated on the last call: 

“We have some production that happens in China and then those components are integrated into bigger components, bigger products at our Thailand facility. And then most of the shipments happen from Thailand, even if they are to the US or to Mexico as well. And then much of the growth that we're seeing is shipments that are coming from Japan as well as from Caswell for the transmission products that we have. So because of that, at least in the short to mid-term, unless policy changes, happen at a government level, we're not expecting much of an impact.” 

This was also stated in the introduction: 

“Second, we are scaling capacity for our highly differentiated laser transmitter chips in our indium phosphide wafer fabs and optical circuit switch and transceiver production capacity in our proven factories outside of China to meet the rising demand.” 

Quite a lot has changed in terms of policy at the government level since the last earnings call. The overall commentary regarding China reliance being minimal on the supply chain side may be true — but tariffs could upend enough of the supply chain to affect Lumentum’s customers (and overall demand). 

Also, despite not relying on manufacturing in China, Lumentum reported Hong Kong as a major customer in the 10-Q at 19% of revenue for LITE, second to United States. Geo-political tensions should cause China/HK to source domestically either voluntarily or through blacklists.  

Source: Lumentum’s 10-QLumentum’s 10-Q 

Telecom Problematic End Market yet DCIs are a Bright Spot 

There has been a steep inventory correction in telecom that has led to substantial revenue decline and significant margin erosion, presenting a major fundamental headwind for data center growth to overcome. However, data center interconnects (DCIs) are helping to drive the turnaround in this otherwise problematic end market.  

Lumentum offers tunable laser and coherent pluggable transceivers for data center interconnects (DCIs). These long-distance data transmissions are traditionally used for telecom purposes and can range up to hundreds of kilometers yet are now seeing demand for data center buildouts.  

In April 2025, Lumentum announced the sampling of new 400/800G ZR+ L-band pluggable transceivers and the general availability of its 800G ZR+ C-band module. The L-band modules effectively double the usable wavelength range and available fiber capacity. By expanding into both C-band and L-band spectrums, Lumentum’s transceivers enable increased fiber capacity, and are able to serve AI and cloud-based applications. These pluggable modules reduce overall system complexity and cost.   

Regarding DCIs, per the previous earnings call: “we're seeing dramatic strength in anything ZR, anything to connect data centers as data centers are being built out, and that can take the form of ZR modules. But given our share of tunable lasers that go into ZRs, that's where we're going to see a dramatic pickup in the telecom side.” 

The most recent update from management is that demand remains very, very strong: “And so the data center interconnect and the networks that are interconnecting these data centers is showing very, very strong demand. And so DCIs and the components as well as ROADMs amplifiers and longer haul coherent transmission connecting further apart data centers is very strong.” 

Revenue Growth Poised to Accelerate  

Q2 FY25 revenue grew by 9.65% YoY and 19.37% QoQ to $402.2 million, driven by strength in its Cloud and Networking segment, beating estimates by 2.87%. 

  • Management guided Q3 revenue between $410 million to $425 million, with a midpoint of $417.5 million for 13.9% growth. 
  • Analysts expect revenue to grow 46.6% in the June quarter to $452 million and 43% YoY in the September quarter to $481.9 million in Q1 F2026. 
  • Management also reaffirmed its commitment to reaching quarterly revenue of $500 million by the end of calendar year 2025, driven by improving trends with its networking equipment manufacturing customers. 

Segments: 

Cloud and Networking Drives Revenue and Margins

In the most recent quarter, Cloud and Networking grew 18.3% YoY and 20.2% QoQ to $339.2 million with management stating they saw “sequential increases in nearly all of our Cloud and Networking product lines.” Cloud and Networking segment profit grew 16.2% for an increase of 330 basis points sequentially and an increase of 610 basis points year-on-year on higher revenue. This segment includes optical transceivers, the datacom chips/lasers that go into optical modules (EMLs, VSCELs, CW lasers, etc) and telecom/data center interconnects. 

Previously, in the September quarter, the segment saw a meaningful inflection with growth of 23% YoY and 11% QoQ. Next quarter, management stated they expect the segment to increase by $25 million QoQ, which would represent growth of 7.3% QoQ – not quite as high as the previous quarter yet it’s been clearly noted in our previous analysis that H2 would be the bigger ramp. Our analysis on delayed Nvidia suppliers also connects some dots on this particular timing for Lumentum. 

According to management commentary, the company shipped record EML units with datacom transceivers shipping to their largest hyperscaler customer and volume production shipments to a new customer.  

The company stated they expect to increase their market share: 

 “Based on the breadth of our 200G EML design wins, we expect to gain additional laser transmitter market share in the upcoming wave of 800G and 1.6T transceivers, utilizing the more efficient 200G EMLs for AI applications. Complementing our EMLs are our new 200G lens integrated photodetector arrays, which adds to our content opportunity in next generation 800G and 1.6T transceivers as well as strengthens our vertical integration strategy for our own cloud modules.” 

Industrial Tech Continues to Contract 

Q2 F2024 Industrial Tech revenue fell (21.4%) YoY and grew 15.4% QoQ to $63 million. The sequential increase was driven by higher industrial laser shipments, partially offset by lower 3D sensing shipments. 

Next quarter, management expects revenues to decline sequentially by $10 million, driven by declines in both commercial lasers and 3D sensing. 

Adj. Margins Bottomed Out and Accelerating  

GAAP Gross margin was 24.7% with adjusted gross margin at 32.3%. Per management, company gross margins will “sequentially increase as manufacturing utilization improves as well as an increase in Datacom laser shipments.” 

  • GAAP operating margin was (12.8%) for operating losses of ($51.6) million. 
  • Adjusted operating margins rose to 7.9% compared to 3% last quarter. It’s expected to further expand to 10% at the midpoint this quarter. 
  • GAAP net margin was (15.1%) for net losses of $61 million 
  • GAAP adjusted net margin was 7.46% for adjusted profits of $30 million 

Adj. EPS Returns to Growth and Accelerates Quickly 

Q2 FY25 adj. EPS improved to $0.42, beating consensus estimates by 16.97%. Management guided Q3 F2025 adj. EPS between $0.47 to $0.53, with a midpoint of $0.53. There is outsized growth in EPS from the rebound on small numbers: 

  • Analysts expect Q4 adj. EPS to grow 957.38% to $0.63  
  • Q1 is expected to see 332.24% to $0.78 

The difference between GAAP and non-GAAP operating margin is due to the stock-based compensation expenses and amortization of acquired intangibles.  

Elevated Capex Spend in Thailand Manufacturing Site Pressuring Free Cash Flow 

Q2 FY25 operating cash flow was $24.3 million or 6.1% of revenue. The Company spent $74 million capex in Q1 FY25 and $64 million in Q2, resulting in sequential negative free cash flow.  

Free cash flow in Q2 was ($15.9 million) or -4% of revenue.   The Company closed the quarter with $896.7 million in cash and cash equivalents and $2.47 billion in debt. This puts the debt-to-equity ratio at 2.75 due to high capex spending — which is high and not ideal in this environment. However, if Lumentum can prove the capex will quickly be converted to revenue, it may become a non-issue by this time next year.  

Lumentum spent $64 million in capex for expanding cleanroom capacity at the Thailand manufacturing site and increasing equipment capacity for indium phosphate wafer production to support EML chip manufacturing. 

Valuation  

Lumentum trades at a forward price-earnings (P/E) of 31.3 and a current PE ratio of 143 (although this looks drastic given the weak bottom line the company is rebounding from). 

The price/sales (P/S) ratio is 2.6 and forward P/S is 2.3. The five-year average P/S ratio is 3.2 

These valuations do not reflect an AI story should Lumentum catch the AI bid (in a bull market) the valuation could be 5-6. This may not be in the near-term given tariff related concerns, weak semiconductor sector performance and Nasdaq entering a bear market officially following the (20%) decline. However, looking into the second half of the year, should conditions improve, Lumentum is capable of trading higher. 

Q&A from Earnings Call 

Yield/Supply Issues are Limiting Growth 

According to the earnings call, the Cloud and Networking segment is expected to increase by $25 million yet the company has “demand that far surpasses that.” Management went on to explain: “we could have probably seen a double-digit increase sequentially quarter-over-quarter had we not had some of those supply chain shortages.” 

It was later more specifically called out as yield issues on the transceiver side, with this comment being in context of the lower margin: “So, yes, so during the quarter, we had some yield issues related to new product ramps within our Transceiver business. That probably was a headwind of anywhere from 100 to 150 basis points.” 

This isn’t exactly surprising given the clear commentary around capex and the need to increase capacity. It was also later stated the supply issues should ease by the June quarter:  

“To your second question, we're gating our module customers’ ability to grow their Transceiver business by the lack of worldwide indium phosphide for EMLs and CW lasers, quite frankly, we're having challenges actually getting enough CW lasers for our own transceivers. So that is actually impacting us on the short-term this quarter, hope to have that resolved in the June quarter. “ 

With even further questioning, it was revealed that hermetic packaging is also creating supply shortages, which refers to sealed enclosures that protect optical components. 

“But the worldwide shortage of things like hermetic packages is creating a challenge for the kinds of volumes that our customers are looking for especially in coherent components and narrow line with lasers. And so, if we have those packages and we could get them more readily, we could grow, as Wajid said, double-digits quarter-on-quarter. It is going to hamper us in both the March quarter as well as the June quarter, and we're working diligently to minimize that impact in the June quarter, but the March quarter is what it is because we need those deliveries now in order to turn products for the quarter and we have a limited supply and ability to get that for example.” 

Hyperscaler Customers 

In the most recent quarter, the 10-Q states that three customers accounted for 16%, 14% and 11% of total revenue with two customers accounting for gross accounts receivable.  

  • In the prior quarter, two customers accounted for 15% and 12% of revenue, respectively.  
  • When comparing to the quarter a year ago, there were three customers at 19%, 13% and 11% of total revenue.  

Management provided the following color in regard to how the ramp is going with the three customers: 

“In Q2, Datacom transceiver revenue grew sequentially as expected, driven by an increase in shipments to our largest cloud hyperscale customer and the start of volume production shipments to one of our new customers we highlighted on prior calls. 

We continue qualification work with the other new customers and expect to start initial volume production during the fourth quarter continuing to ramp through the first half of fiscal '26. Transceiver manufacturing capacity expansion is also progressing as planned.” 

During the call, there was a Q&A exchange that drilled deeper into a potential fourth customer: 

“Ananda Baruah:  

 “On the new ramping customer, can you give us some sense of time frame when you think that it hits run rate and or any context around what run rate I know this is probably kind of somewhat project based, but when you think it hits run rate, what time frame? And then the second one is sounds like this year, calendar year '25, you're not going to run into sort of congestion between your EML chips and your own transceivers since they're largely SiPho (silicon photonics) based right now.”  

Lowe:  

 “As far as the new customer reaching run rate, I'd say that's going to take some time. So don't be raising your projections for that customer. But as I said earlier, we're qualifying a second product there that should come on by the end of the calendar year. So, I'd say around the end of the calendar year, we should be in full motion with that customer. And then as we talked about in the script, the third customer start production in fiscal Q4 and really hit run rate, I'd say, by the fall time. So, September, October. That product is scheduled to ramp significantly faster. So that's the ramp color.”  

Conclusion: 

Lumentum is a small-cap company with differentiated technology that has caught the attention of heavyweight NVIDIA. We’ve been closely monitoring Lumentum for roughly a year, waiting patiently for their EML lasers for 200G to ship, enabling 800G and 1.6T bandwidths. Any progress here should continue into 2026-2027 for 400G data lanes and 3.2T bandwidths. 

Their indium phosphide (InP) laser technology offers significant power efficiency advantages over traditional silicon photonics. This is increasingly important as power consumption becomes a central concern in scaling AI data centers. Lumentum’s collaboration with NVIDIA, integrating their high-efficiency lasers into NVIDIA’s Spectrum-X and Quantum-X photonics networking switches, is a nod towards the EML lasers and their ultra high power lasers for CPOs playing a critical role in future architectures. 

With that said, Lumentum is in the high risk bucket due to being a small cap. The stock requires speculation as to when a shift in fundamentals will occur. To date, we have seen an inflection for one quarter, yet we need more evidence before an inflection becomes a meaningful trend. The sharp acceleration provided for in analyst consensus in H2 could wane if supply chain troubles trickle down and result in slower sourcing for AI systems.  

Ultimately, it’s well worth our time to earmark companies like Lumentum and identify potential entry targets.  

The I/O Fund recently launched our new Discovery tier which surfaces new ideas the I/O Fund does not own yet at a pace of 30-40 new stocks per year. Coverage will include AI hardware, AI software, crypto and more, from a leading tech portfolio. new Discovery tier which surfaces new ideas the I/O Fund does not own yet at a pace of 30-40 new stocks per year. Coverage will include AI hardware, AI software, crypto and more, from a leading tech portfolio.  

Sample research we published in March and April:   

  • Key supplier to TSMC’s new high-growth platform called Compact Universal Photonic Engine (COUPE)  
  • Nuclear and natural gas supplier for AI data centers   
  • A breakdown of the risks and opportunities for the biggest IPO in the AI sector

Please note: The I/O Fund conducts research and draws conclusions for the Fund’s positions. We then share that information with our readers. This is not a guarantee of a stock’s performance. Please consult your personal financial advisor before buying any stock in the companies mentioned in this analysis.

Recommended Reading:

The Impact of Tariffs on the Stock Market: Q1 Preview

Posted on by io-fund

Heightened volatility and extreme news-driven sentiment in the markets are leading to quick whipsaws of 3% or more in either direction daily. These swings of quick drops to quick pops are getting traders addicted to the idea that tariff-related macro uncertainties will resolve smoothly in the end. 

Despite the markets approaching bear market territory in March, the market is rallying on the hope that geopolitical tensions will ease. Meanwhile, analysts are revising estimates under the hood with cautious notes that these issues will not disappear overnight. The reality of what it will take to move supply chains will eventually settle in. 

Below, we dig into early commentary from executive teams and analyst observations in early March and April pointing to growing uncertainty on customer behavior and demand, supply chain challenges, and how this sets the stage for Q1’s earnings season and beyond. 

Consumer Electronics in Focus as Analysts Estimate Apple’s Impacts 

Apple will be a central player in the potential tariff impacts, being nearly a pureplay in the consumer electronics sector with heavy reliance on China. 

Apple’s diverse global supply chain exposes it to multiple tariffs of different rates depending on production location, with iPhone production concentrated in China but also located in India and Brazil. Apple is said to have quickly shipped 600 tons of iPhones, or ~1.5 million worth nearly $2 billion, from India before tariffs were implemented, while also increasing production of the iPhone 16e in Brazil to avoid higher Chinese tariffs, 

As to how tariffs could impact Apple and its prices, analysts broadly see higher prices on the horizon, regardless of whether Apple passes tariffs on or works to bring manufacturing to the US to avoid paying tariffs.  

Wedbush analysts estimate that onshoring iPhone production to the US could increase iPhone prices to $3,500, assuming it would cost $30 billion over 3 years to bring just 10% of its supply chain onshore.  

BofA estimated that based on higher labor costs alone, iPhone 16 Pro prices could rise at least 25%, from $1,199 to $1,500 — this does not even account for higher chip prices or higher component prices.  

In a report from earlier this week, Citi estimated that iPhone prices would rise 7% globally “if China/India total tariffs are 45%/10%, assuming another 25% tariff on Apple products exported from China to the US through Section 232, and Apple (passing) all tariff costs to end customers.”  

However, on Wednesday morning, the administration floated the idea of China tariffs to a baseline 50%-65%, above Citi’s estimates and likely to further inflate prices.  

Other analysts see much higher prices even if Apple passes on the entirety of the tariff burden to consumers — Rosenblatt estimates that iPhones could be up to 43% more expensive, while Counterpoint Research estimates prices could be up to 30% more depending on manufacturing location.  

Evercore estimates Apple could face a $9-10 billion impact to COGS at an effective tariff rate of 16%, which they expect would impact full-year EPS by  (7%), or $0.51.  

Price hikes to this degree could negatively shock demand, with the smartphone market already on thin footing this year. Global smartphone shipments rose 1.5% YoY in Q1, per IDC, though the group said this was due to a supply-side surge in shipments ahead of tariffs. 

IDC said this dynamic “effectively inflated Q1 shipment figures beyond levels anticipated based on underlying consumer demand trends alone,” adding that heightened geopolitical tensions between the US and China and growing tariff uncertainties were a “strong reason for concern” for 2025 growth.  

TrendForce estimated that the “best case scenario will see the smartphone market flat at best” in 2025, while the “worst case scenario is a production decline by as much as 5% YoY.”  

The PC market is in a similar situation as smartphones, with shipments rising in the high single-digits in Q1, again with the increase driven by vendors stockpiling rather than strong consumer demand.  

IDC recently cut its PC forecast for the year, seeing 3.7% growth in global shipments and just 0.2% growth in consumer shipments, as tariff-driven price hikes “combined with subdued demand are leading to a negative impact within the largest market for PCs.” Canalys expects that “subsequent quarters this year are likely to see a slowdown as inventory levels normalize and customers face higher prices,” even as the Windows 10 end-of-life and upgrade cycle looms.

Auto Facing Rapidly Shifting Tariff Policy, Uncertainty 

Auto executives united this week to lobby against the 25% tariff the sector faces, saying the tariffs will upend the global supply chain and cause a domino effect of higher prices for consumers, lower vehicle sales and higher repair costs. On April 23, the administration noted that Trump was considering exempting auto parts from tariffs on China imports, though the separate 25% tariff would remain and go into effect on May 3.  

Ford already warned dealers that it anticipates increasing prices for May production, after halting shipments of US-made vehicles to China due to the high tariffs. Notably, Tesla walked back on its growth guidance for 2025 this week, while auto-exposed semi Aehr warned of uncertain customer behavior and demand.   

Aehr, Tesla Pull Guidance on Tariff Uncertainty 

Reporting shortly after tariffs were announced at the start of April, semi small-cap Aehr provided some insight into the auto market, an important first look considering its high customer concentration with Onsemi.  

Aehr’s management believes the company will not face significant impacts from tariffs, though the main concern they echoed was that the real challenge “is not being able to control near-term secondary effects on our current and potential new customers, such as possible near-term delays in customer orders or requested delivery dates.” Because of the broader impacts to the customer ordering behavior, supply chain and shipment delays, and tariff implementation, Aehr pulled its guidance and stated they would re-evaluate as they get more clarity. 

On Tuesday, Tesla also pulled its guidance in similar fashion, stating that it is “difficult to measure the impacts of shifting global trade policy on the automotive and energy supply chains, our cost structure and demand for durable goods and related services.” Tesla also shifted its tone on vehicle growth for 2025, first saying in Q4 that it expects to return to growth in 2025, but now saying the growth outlook would be revisited in Q2. A company statement added that the rapidly changing policy “could have a meaningful impact on demand for our products in the near term.” 

Significant Downward Revisions Hit Tesla After Q1 

Tesla sits at the crossroads of consumer demand and China tensions. The company’s Q1 report revealed margin weakness once again, along with a large revenue and EPS miss. Tesla’s estimates for 2025 have already faced substantial pressure over the last four months, and that’s before additional industry-wide growth concerns rose with tariffs. 

Automotive gross margin (excluding regulatory credits) declined more than 1 point QoQ to 12.5%, while operating margin dropped to just 2.1%. As a result of the nearly $2 billion revenue miss and contracting margins, adjusted EPS declined (40%) YoY to just $0.27, missing the $0.42 consensus by a wide margin. This weighed heavily on growth expectations for 2025, as EPS generation was expected to improve significantly through the rest of the year. 

In just the first day after Q1’s report, Tesla already witnessed a major haircut to growth expectations for both revenue and EPS. 2025 revenue was revised $10 billion lower to $96.8 billion, for a (1%) YoY decline, while EPS was revised more than $0.50 lower to $1.98, for a (20%) YoY decline. 

Graph of Tesla's 2025 revenue and EPS estimates showing sharp downward revisions after Q1 2025's earnings report. Source: YCharts

Tesla’s 2025 revenue and EPS estimates were already under pressure, but the stock saw additional negative revisions after Q1’s double miss. Source: YChartsYCharts 

Consider that at the start of 2025, Tesla was expected to see nearly 20% YoY revenue growth and 30% YoY EPS growth. After just the first quarter, Tesla’s revenue has been revised a total of $20 billion lower and EPS more than $1.20 lower, marking a significant erosion of growth even before tariff impacts on costs and demand are felt.  

The I/O Fund specializes in covering lesser-known AI stocks on our research site with trade alerts and weekly webinars. Learn more here.The I/O Fund specializes in covering lesser-known AI stocks on our research site with trade alerts and weekly webinars. Learn more here. 

Indirect Tariff Impacts Arise for Big Tech  

From Big Tech, we’re seeing some indirect tariff impacts arise, notably from major Chinese retailers pulling ad spending on Meta and Google. 

PDD’s Temu and Shein both are drastically reducing ad spending in the US, after the end of the ‘de minimis’ exemption on packages from China valued at less than $800 earlier this month. Temu reportedly “axed all spending on Google’s Shopping platform since April 9,” while data showed that its daily ads run on Google plunged 99.9% to just 14, down from 30,000 to 60,000 at the beginning of April. 

Across popular social media platforms including X, YouTube, and Meta’s platforms, Temu reduced spending “by an average of 31% in the two weeks leading to April 13 compared with the previous month,” per SensorTower. Shein’s daily ad spending “across Meta, TikTok, YouTube and Pinterest fell 19%” over the same period. However, SensorTower noted that Temu’s ad spend had risen so quickly, that even after the reductions, it was still above 2024’s level.  

For Meta, China accounted for ~11% of revenue in 2024 at $18.4 billion, while for Google, YouTube accounted for more than 10% of revenue at $36.1 billion. Lost spending from two popular retailers could potentially lead to billion-dollar revenue impacts that then weigh on EPS.  

Over the past month, Meta has seen 2025 revenue estimates move more than $2 billion lower to $186.4 billion, and EPS estimates (2%) lower to $24.90, now pointing to growth of just 4.4% YoY. Google’s impacts are more limited at this time, with revenue and EPS each revisions under (1%). 

Graph of Meta and Alphabet stocks seeing negative revisions to revenue and EPS in April after tariffs implemented. Source: YCharts

Since tariffs were implemented in early April, Meta and Alphabet stocks have seen negative revenue and EPS revisions. Source: YChartsYCharts 

AI Semis, Hardware 

Recent reports from across the AI hardware industry, from semis to servers, point to an uncertain environment stemming from tariffs …

TSMC Sticks to Mid-20% Sales Growth Guide on Strong AI Demand 

TSMC stuck to its 2025 forecast for mid-20% sales growth, as it guided Q2 revenue growth ahead of consensus at 38% YoY after beating on the top and bottom line in Q1. Management addressed tariff impacts and acknowledged uncertainties, noting that there might be more clarity in a few months’ time. However, there are signs that tariff-related effects and higher costs will flow downstream to TSMC’s leading customers such as Apple.  

CEO C.C. Wei seemed to brush off broader growth concerns, stating that the chipmaker has not “seen any change in our customers behavior so far” with AI demand remaining robust throughout 2025. Wei acknowledged that TSMC “might get a better picture in the next few months” on tariff impacts to end market demand.  

Despite management’s confidence in riding strong AI demand to mid-20% growth this year, analysts are growing more concerned about the full-year guide given the risks key customers are facing. JPMorgan analysts say TSMC “could pare [its forecast] slightly to target low- to mid-20%” sales growth, while Deutsche Bank analysts raised the concern that the chipmaker “may also withdraw its guidance as customers adjust to tariffs.”  

Estimates have yet to reflect some of these concerns, with 2025 revenue forecast at $113.7 billion, up from $111.2 billion at the end of March, likely due to Q2’s guidance beat. EPS estimates are also up 2% over the past month to $9.46, or 31.5% growth.  

On a quarterly basis, Q4 is expected to be the weakest, with revenue growth decelerating to 11.5%. Q4’s revenue has been revised nearly (6%) lower over the past three months and (4%) lower over the past month to $29.4 billion. EPS mirrored this, revised (6%) lower over the past three months and (5%) lower over the past month to $2.57, for growth of just 5.5%. This raises some initial red flags about holiday demand for consumer electronics and other devices. 

Chart of TSMC's quarterly EPS revisions showing positive EPS revisions through Q3 and negative revisions in Q4. Source: Seeking Alpha

TSMC’s Q4 earnings have been revised nearly (6%) lower over the past three months in stark contrast to revisions in Q2 and Q3. Source: Seeking Alpha 

TSMC’s US production push could also impact customers alongside tariffs. Recently, TSMC boosted its US investments to a total of $165 billion in its Arizona complex, now aiming to construct six fabs and an AI research facility. The first fab is in volume production for the 4nm node, while the second fab for 3nm chips completed construction with TSMC aiming to accelerate volume production to meet demand – Nvidia and AMD both announced last week intentions to ramp production with TSMC in the US as tariffs loom. 

While ramping US production theoretically would eliminate tariff risks, it still could have downstream effects on end market customers and consumers that mimic higher prices that tariffs bring. Reports suggest that TSMC is considering hiking 4nm prices by up to 30% in the US, as primary customers Apple, AMD and Nvidia are said to be rushing in orders. Other reports suggest that all of TSMC’s US chips could see prices hikes of at least 15% due to higher labor costs and depreciation. It’s unlikely TSMC’s top customers will be willing to simply absorb a 30% increase in US-made wafer prices without passing some or all of these costs on, which could add more price pressure on top of tariffs for electronics products such as smartphones, PCs, gaming chips, and more.  

ASML Says Tariff Impacts Uncertain, Stands by 2025 Guidance 

ASML also stood by its 2025 sales guidance despite missing rather widely on bookings, and its management team was much more cautious than TSMC on the broader macro impact from tariffs, stating bluntly that “it is clear that uncertainty is increasing in the macro environment.” 

ASML’s management explained that while tariff discussions were just starting, the “end state will be unknown for a while and until then, the potential impact” will remain unclear, and gross margin impacts would be “absolutely impossible” to quantify. CEO Christophe Fouquet added that uncertainty at some customers could push 2025 sales towards the lower end of its €30 billion and 35 billion range, though strong AI demand could push it towards the upper end, hesitating to commit to either side.  

Although ASML did not quantify potential impacts, a Reuters report outlined impacts to US-based WFE makers. Reuters noted that according to industry calculations shared with lawmakers, tariffs could cost the US WFE industry $1 billion annually, with Applied Materials, Lam Research and KLA all facing impacts of up to $350 million each.  

Expectations for ASML have strengthened over the past three months, even with bookings missing estimates by nearly €1 billion and Q2’s guide coming in soft. Revenue estimates through Q4 have been revised 7% to 14% higher over the past three months, while EPS estimates have risen 1.7% to 4.4%. For the full year, both revenue and EPS have been revised 11.5% to 12.5% higher.  

ASML stock's revenue and EPS have seen large positive revisions in the 11% to 13% range despite bookings missing estimates in Q1. Source: YCharts

ASML’s revenue and EPS have both been revised more than 10% higher since the start of 2025 despite Q1 bookings missing estimates by nearly 1 billion. Source: YChartsYCharts 

Management signaled an intent to pass tariff costs on and not bear any of the burden themselves, while questioning how tariffs would “ultimately be absorbed in the entire value chain.” ASML said that they are working to “minimize the total exposure of the ecosystem to tariffs,” but once it has been minimized, they will pass the tariff burden on to the “next element in the value chain.” While ASML is exempt from tariffs at the moment, parts and inputs such as steel and aluminum are not. With ASML shipping parts, inputs and tools up to “multiple times” between the US and Europe, tariff impacts could quickly add up. 

Thus, it’s likely that TSMC, Intel and Samsung, could face rising costs for ASML’s machines, especially considering that the blended ASP for ASML’s low-NA EUV machines sits at 227 million euros, or ~$258.8 million. For TSMC’s case, analysts estimate that ~65% of its $100 billion US investment announced in March could go to WFE, and if it faces 15% higher costs from tariff impacts, that could mean an additional $6 billion more it would have to spend for the same equipment, and an additional $6 billion to pass on to customers to maintain margins.  

Micron Not Including Tariffs in Guidance, Estimates Dropping 

On the memory side, Micron has flagged two key factors: it intends to pass along any tariff costs to customers, and that it has not included impacts from “potential new tariffs” in its guidance offered in March due to a lack of clarity on tariff timing and implementation. Micron noted that it “serves as the U.S. importer of record for a very limited volume of products that would be subject to newly announced tariffs on Canada, Mexico and China.” 

For its fiscal Q3, Micron had guided for record revenue of $8.8 billion, +/- $300 million, ahead of estimates, though it projected gross margin to decline 130 bp sequentially to 35.5%, in part due to higher product mix and weaker pricing in consumer-oriented products. Guidance also included $1.13 billion in operating expenses, and growth in DRAM and NAND bit shipments. However, recent industry forecasts suggest that memory chip demand is disrupting typical seasonal trends and has been “largely frontloaded into the first half of 2025” as US-based firms work to beat tariffs.  

Being more heavily exposed to the dynamic consumer markets opens Micron up to more risk than say a company like ASML, as not only does Micron have to contend with some tariffs on its products, but also more directly with end market demand, which could struggle if smartphone or PC prices rise and dent demand.  

Management previously expected PC growth to be weighted toward the second half of 2025, and smartphone growth in the mid-single digits. As noted previously, industry estimates are pointing to a much more challenged growth picture for the two markets. With Micron seeing 30% revenue exposure to PC, graphics, and mobile end markets, slower end market growth than management currently expects can weigh on Micron’s 2H revenue growth. 

Should tariff related price and demand impacts pressure margins, Micron could see further downside to EPS, which has already come down more than (20%) over the last five months. In early December 2024, Micron was estimated to generate almost $9 in EPS in fiscal 2025, but after its weak Q2 guide, that has now come down to $6.9. A (10%) impact from tariffs could take Micron from the $7 level to the low-$6 range.  

Graph of Micron stock's fiscal 2025 EPS estimates dropping from ~$9 to below $7 after its weak Q2 guide, before any tariff impacts. Source: YCharts

Micron’s fiscal 2025 EPS has already been revised nearly $2 lower after its weak Q2 guide, before tariff impacts are felt. Source: YChartsYCharts 

On a quarterly view, revenue estimates through Q1 2026 (Nov 2025 quarter) have been largely unchanged over the last three months, though EPS revisions range from (2%) to over (6%) lower likely on margin risks. 

The I/O Fund specializes in covering lesser-known AI stocks on our research site with trade alerts and weekly webinars. Learn more here.The I/O Fund specializes in covering lesser-known AI stocks on our research site with trade alerts and weekly webinars. Learn more here. 

HPE To Mitigate Tariff Impacts, Sees ~4% Impact to EPS 

HPE was one of the only AI-exposed firms so far to quantify potential EPS impacts from tariffs in early March, though it is facing broader pressures on margins and EPS from higher AI server inventories.  

For Q2, HPE guided well below estimates for adjusted EPS, seeing $0.28 to $0.34 versus consensus at $0.50. HPE also cut its forecast for the full year, seeing adjusted EPS between $1.70 to $1.90, compared to consensus at $2.13.  

This was driven by issues in HPE’s Server segment — management said that Server margins were adversely impacted by higher discounts, aggressive pricing competition, and “higher-than-normal AI inventory caused by the rapid transition of demand to next-generation GPUs and related components.” Though HPE took action to mitigate these impacts, it noted that these headwinds will apply continued margin pressure over the next one to two quarters.  

Additionally, HPE placed a number on tariff impacts, adding that the majority of it would hit in Q2:  

“What we've got in the guide is actually $0.07 for the year and that gets you to the midpoint of the $1.80 on the guide. I would add though the way to think about tariffs is, we've worked obviously on mitigation effects, but it does take time to see those mitigation effects take place. So as a result, literally $0.04 of that $0.07 is actually going to be in Q2, and it's also in the Server business which is where we see the greatest extent of that [tariff] impact.” 

This would be just nearly (4%) impact to EPS, though management shared broader concerns over how tariffs would impact end market demand in the second half of the year. If tariffs indeed lead to lower demand, higher costs and more prolonged margin headwinds, it could result in a much larger EPS impact given 60% of the $1.80 guide is weighted in 2H. For example, if tariffs have closer to an 8% to 12% adverse impact to EPS, to the high $1.60 range, HPE could be looking at a YoY decline of (14%) to (17%) versus the (10%) currently expected.  

Graph of HPE stock's fiscal 2025 EPS and EPS growth estimates showing shift to negative growth and large EPS cut in Q1. Source: YCharts

HPE’s fiscal 2025 earnings growth estimate has dropped from 10.5% YoY to (6.8%) YoY as a result of its cut guide, with tariff impacts yet to be felt. Source: YChartsYCharts 

What This Means for Q1 and 2025 

It’s no doubt that April has been quite a rocky month for stocks as earnings season ramps up, with the Dow headed for its worst April return since 1932, the VIX reaching levels higher than in all of 2022, and bearish sentiment reaching extremes, as we outlined in our analysis The Fed Can’t Save This One: Why Bonds May Break the Stock Market in 2025.  

Initial commentary from Q1’s first reports and broader macro developments in early March and April raises a few major issues for the Q1 season and for 2025: 

  1. Increased macro uncertainty weighing on customer behavior and impacting ability to forecast growth, with numerous companies pulling guidance 
  2. Supply chain challenges and higher prices as tariffs rise 
  3. Earnings and revenue expectations being revised lower, though full tariff impacts are yet to be felt 

Tariffs could quickly complicate the global supply chain and have trickle-down effects to consumers, as it’s impossible to onshore complex supply chains to the US overnight, or in short order, without facing major increases in costs. Companies like Apple may also be hesitant to commit to onshoring significant levels of manufacturing given the billions in costs and years it would take. Companies must also consider the fluctuating trade policies could shift in a way that would make domestic production unfavorable in the future.  

The auto and consumer electronics industries indicate that rising part and component costs from tariffs can and likely will lead to higher product prices in the coming months/quarters, which may then in turn pressure consumer demand. Should weaker consumer demand lead to production cuts or rising inventories, that could exacerbate margin pressures and lead to further downside to EPS. 

As of now, there’s been limited commentary about the potential impact of tariffs on growth and earnings, with most tech companies simply saying the environment is too uncertain to forecast or indicating an intent to pass costs along the chain to customers and thus landing with consumers.  

This begs the question — what if tariffs cause much larger negative impacts to EPS this year, and in turn, what do growth expectations then look like? HPE was one of the few so far to put an EPS impact from tariffs at approximately (4%), but if the impact is closer to (10%) or even (15%) from higher costs and weaker demand, corporate earnings growth expectations would need to be revised much lower.  

Similar to what we discussed last week in our analysis Tesla Stock Faces Recalibration of Growth Expectations, downward revisions could easily snowball and take growth from high double-digits to single-digits, from up single-digits to down single-digits, or from down single-digits to down double-digits.  

What this Means for Tech Valuations 

What’s most important for investors is what all this uncertainty means for tech valuations. Social media would have you believe the world is ending one day, and that a China deal would make all these issues simply disappear with no lasting impact to be felt.  

On the semi and AI hardware side, valuations are reaching multi-year lows, with TSM trading at its lowest forward P/E since the start of 2023 and ASML trading at its lowest since early 2020.  

Graph of TSM, ASML stock forward P/E ratios since 2023 showing multiples at multi-year lows. Source: YCharts

On a forward P/E basis, TSM and ASML are trading at levels not seen in at least a year. Source: YChartsYCharts 

Ahead of ASML’s earnings report last week, there was rising talk and belief that the worst expectations and many risks are being priced in, as its forward P/E had compressed by more than half since its peak above 50x in July 2024.  

However, we have yet to fully understand and quantify the impacts of tariffs to revenue, margins, EPS and customer demand over the next few quarters. There’s risk that growth expectations get re-rated lower for leading tech stocks, either via direct tariff impacts or indirect impacts from lowered spending, macro slowdowns and/or weaker consumer demand.  

For most of the Mag 7, investors have been accustomed to strong earnings growth since 2023 as the tech giants captured AI tailwinds; for 2025, Meta, Alphabet and Amazon are facing significant decelerations in EPS growth, in part due to tough comps from high growth last year.  

Table showing Apple, Meta, Alphabet, Tesla and Amazon stocks' historical EPS growth from 2023 and 2024 and forecast EPS growth through 2025

Some of the Mag 7 leaders over the past two years are expected to see EPS growth dramatically decelerate in 2025. Source: YCharts YCharts  

For Meta and Google, should lower ad spending from major Chinese clients, and weaker economic growth weigh further on ad spending, both could see increased risks from losing multiple-billions in high-margin ad revenue, considering it flows heavily to the bottom line.  

Graph of Amazon, Meta, and Alphabet stocks' forward P/E ratios since start of 2023. Source: YCharts

Alphabet, Amazon and Meta are trading near the lowest forward P/E multiples since early 2023. Source: YChartsYCharts 

Investors are still paying 30x forward earnings for Amazon despite headwinds to retail sales and possible risks to cloud consumption growth and ad spending.  

Valuations may be approaching lower levels, but the risks on the horizon have amplified — for the tech sector, macro headwinds are becoming much more pronounced, with risks to customer demand on core growth drivers, whether that be cloud, ads, auto or consumer electronics.  

Conclusion 

Analysts are beginning to revise earnings and revenue estimates lower as uncertainties rise, and this will likely continue throughout Q1 and into Q2 due to the effects of tariffs.  

Investors have been trained to buy-the-dip, and to expect the dip will always quickly recover in short order; however, as we have cautioned earlier this month, investors need to be prepared for a changing dynamic as bonds are equally as concerning in terms of what could impact the market this year. 

If you want to identify which Mag 7 stock is the strongest, and which stocks are the weakest in this new tariff-driven economy, then we encourage you to attend our upcoming weekly webinar for premium members. Join us Thursdays at 4:30 p.m. EST to hear our game plan regarding the remainder of 2025 including our strategy to raise cash and further hedge. Our cumulative returns of 210% and annualized returns of 27.6% place us as one of the top performing tech portfolios, beating Wall Street’s very best. Learn more here.

Please note: The I/O Fund conducts research and draws conclusions for the Fund’s positions. We then share that information with our readers. This is not a guarantee of a stock’s performance. Please consult your personal financial advisor before buying any stock in the companies mentioned in this analysis.

Recommended Reading:

Amazon: AI Powerhouse Driving Margin Transformation, Retail in Tariff Crosshairs

Posted on by io-fund

Amazon is building an AI powerhouse in AWS, with the segment becoming an increasingly large driver of Amazon’s earnings power despite contributing just one-sixth of Amazon’s revenue. Management remains committed to continuously expanding AI offerings and expanding GPU capacity to meet demand, seeing long-term tailwinds to its AI growth story. 

E-commerce remains Amazon’s bread-and-butter, but it has now found itself caught in the crossfire of fluctuating tariff policy, being quite heavily exposed to China. Amazon could feel impacts to both revenue and margins as a result, given that peer Walmart pulled its operating income growth forecast of 0.5-2.0% and simply said it has “widened” on tariff uncertainties. For Amazon, retail margins have also improved over the past few quarters to help (minimally) offset the effects. 

AWS has helped drive strong expansion in Amazon’s margins with operating profit up 61% YoY to a record $21.2 billion. With that said, this previously bulletproof segment is also exposed to the trickle-down effects of tariffs as cloud customers exposed to China will likely cut their budgets in response. Mizuho believes this could be up to 50% of cloud customers.  

How much of this is priced in? That’s a question the I/O Fund is working hard to answer as the risks in the market are immense yet shares of Amazon are near historically low levels.  

Building An AI Behemoth in AWS 

Amazon’s approach in AI mirrors a one-stop-shop, where it is providing access to a wide range of the industry’s leading models alongside its internal models, access to the newest generations of GPUs alongside extensive custom silicon deployment, and any service that customers could need in-between. Amazon stated in Q3 2024 that it had launched nearly 2x as many genAI and ML features as the other CSPs combined over the prior 18 months.  

This is what Amazon sees as its differentiator versus Azure or GCP – its ability to offer more AI/ML features and services, a wide range of powerful custom silicon and GPU instances, and end-to-end platforms for developers to build on.  

Amazon’s strategy is centered around offering efficient access to AI on a price-performance basis at any size or scale, either using its custom AI accelerators for training (Trainium) or inference (Inferentia), a wide range of Nvidia GPUs, or custom CPU instances (discussed below).  

Customers can test, build, and customize genAI apps on the latest LLMs from a wide range of leading providers on Amazon’s serverless platform Bedrock, while SageMaker AI is AWS’ fully managed AI/ML platform spanning the training, fine-tuning and deployment lifecycle. Customers also have access to Amazon’s genAI assistant Amazon Q, and a wide range of storage, analytics and security tools.  

AWS has developed several networking innovations to improve performance and scalability for AI workloads, such as Elastic Fabric Adapter (EFA) and Nitro. EFA can communicate directly with networking hardware using a technology called operating-system bypass (OS-bypass), reducing latency and allowing AI and HPC applications to easily scale to thousands of GPU or CPU cores. Nitro offloads typical virtualization functions such as networking, storage, and management tasks to specific Nitro cards, freeing up server resources and reducing costs.  

Breakdown of Amazon’s Custom Chips, Instances  

AI chips are a performance game, with Nvidia quickly upping the ante with its GB200 NVL72 systems, which offer up to 30x faster inference performance on trillion-parameter LLMs. Google and Amazon are responding quickly with their custom accelerators, with Google recently unveiling its 7th-gen TPU, offering up to 10x faster AI processing than v5p. Amazon says that its next-gen Trainium3, due later this year, is 2x as fast and 40% more energy efficient than Trainium2. 

Amazon boasts a wide range of custom accelerator and CPU instance offerings alongside leading Nvidia GPUs:  

Trainium2 AI Accelerator Offers up to 40% More Price Performance 

Amazon launched its Trainium2 chips and instances in November 2023, built primarily for AI training tasks, offering up to 4x the performance of its first generation Trainium chip. Trn2 instances combine 16 Trainium2 chips to offer 20.8 petaflops of dense FP8 compute. Compared to its next-most powerful EC2 instances, Trn2 offers “30% more compute and 25% more high bandwidth memory.” 

Amazon says that the Trn2 are “typically 30% to 40% more price performant than other current GPU-powered instances available.” For example, Databricks uses Trn2 instances to help “lower TCO by up to 30% for its customers,” while DataDog says the instances help users “cut AI infrastructure costs by up to 50% and boost model training and deployment performance.” 

Next-Gen Trainium3 Set to Launch Later in 2025 

Amazon’s next-generation Trainium3 chip is set to debut later in 2025, and are expected to offer up to 4x the performance of Trn2 UltraServers. Amazon has been sparse on details for the new chip, aside from the performance boost, speed upgrade and energy efficiency improvements. AWS CEO Matt Garman said the new clusters will let customers “iterate even faster when building models and deliver superior real-time performance when deploying them.” 

Inferentia Tailored for Fast, Low Cost Inference 

Amazon’s Inferentia chips, first launched in 2019, were tailored specifically for AI inference applications and aimed to tackle the barrier of high inference costs, which at the time could account for up to 90% of the infrastructure cost for building and deploying an AI/ML app. 

The first-generation chip offered up to 2.3x higher throughput and 70% lower cost than comparable EC2 instances on inference tasks. Each chip featured 4 first-generation NeuronCores with 128 TOPS performance for lower precision FP16, BF16, and INT8 calculations, scaling up to 16-chip instances offering 192 GB memory for heavy inference tasks. Inferentia helped Amazon deliver high-throughput, low cost AI inference due to the chip’s inference-optimized design and low instance cost. 

Inferentia2 Provides 10x Lower Latency  

Inferentia2 was built for large-scale AI inference workloads, with Amazon saying customers could effectively deploy 175B parameter models for inference workloads on a single Inf2 instance, which features 16 Inferentia2 chips.  

Inferentia2 can deliver up to 4x higher throughput and up to 10x lower latency compared to Inferentia, as it has 4x more memory capacity and 16.4x higher memory bandwidth than Inferentia. In addition, Inf2 are the first inference-optimized instances to support scale-out distributed inference for large-scale LLMs with high-speed connectivity between chips.  

Inf2 is a core part of Amazon’s strategy to lead in AI inference as inference costs come down, offering low-latency, low-cost inference-optimized performance. Inf2 is said to cost ~$0.40 per 1 million tokens for a 70B model, versus ~$1.00 per 1 million for an H100 GPU ran on Azure. AWS customers can also build across Amazon’s full stack, Trn2 and Inf2, for quick, efficient and cheap AI training and inference tasks. 

Graviton4 CPU Offers 40% Better Price Performance 

AWS’ Graviton CPUs are its in-house Arm-based custom data center chips, with the most recent Graviton4 launched alongside Trainium2 in 2023. Amazon says the next-gen Graviton4 chips can offer up to “nearly 40% better price performance versus other leading x86 processors,” with 30% better performance, 50% more cores and 75% more memory bandwidth than the Graviton3.  

Amazon has built more than 2 million Graviton chips to date since the first generation launched in 2018, with a wide range of instances powered by the chips. AWS offers EC2 instances for general purpose compute, as well as compute optimized, memory optimized, or storage optimized instances. 

Unique ‘Burstable’ T3 Instances 

Amazon’s EC2 T3 instances are a low-cost, burstable instance that offers balanced compute, memory, and network resources primarily for general purpose workloads. What’s unique is that T3 instances provide a baseline level of performance from the CPU and an  ability to ‘burst’ CPU usage to access more compute and reach full-core performance. 

The unique ‘burst’ factor of T3 instances is built upon credits, which are earned when running below baseline or when idle. As needed, at any time, credits can be used to throttle to maximum performance until credits run out, or for as long as required with customers responsible for the additional cost. 

New UltraServers for Scale-out Compute 

UltraServers are a new offering for Amazon, where it is linking four Trn2 instances to form a 64-chip cluster (4 Trn2 instances) for 83.2 petaflops of dense FP8 compute, utilizing its proprietary chip-to-chip interconnect NeuronLink. UltraServers aim to accelerate the training process further, letting customers deploy larger models faster – Amazon says customers could train 300B parameter models in just weeks as opposed to months. 

UltraServers are then linked together to form UltraClusters, letting customers scale-out to tens of thousands up to 100,000 Trainium2 chips. This is the backbone of Amazon’s Project Rainier supercluster project for key partner Anthropic. 

Amazon Spending Same Amount as AWS’ Run Rate on Capex 

Amazon has the benefit of becoming an AI behemoth due to AWS’ massive scale – the unit recently crossed a $115 billion annualized revenue run rate, more than doubling in the past four years, with revenue up 19% YoY to $28.8 billion in Q4. Unlike Microsoft, Amazon has not provided exact AI revenue figures for AWS, stating that it was at a “multi-billion dollar annualized revenue run rate” and growing triple digits YoY in Q4.  

AWS has grown, and continues to grow, significantly quicker than Amazon as a whole. Ten years ago, AWS generated just $4.6 billion in revenue versus $89 billion for Amazon. Since then, AWS revenue has grown nearly 24x while Amazon has not even grown 7.5x.  

AWS is also estimated to have nearly 4.2 million business customers, an enormous base to fuel continued growth in cloud services and AI. However, HG Insights estimated that 92% of these customers spend less than $1,000/month on AWS services, which might be an issue down the line considering that AI’s earliest adopters and largest spenders are likely to be in the Fortune 500.  

Q4’s call also featured an important quote from CEO Andy Jassy about AI’s future at Amazon:  

“I spent a fair bit of time thinking several years out. And while it may be hard for some to fathom a world where virtually every app has generative AI infusing it with inference being a core building block just like compute, storage and database, and most companies having their own agents that accomplish various tasks and interact with one another. This is the world we're thinking about all the time and we continue to believe that this world will mostly be built on top of the cloud with the largest portion of it on AWS.” 

Amazon is buying fully into this vision, indicating that it is spending nearly AWS’ revenue amount in capex in 2025 to capture growing AI demand. Jassy explained that annualizing Q4 2024’s $26.3 billion capex would be a reasonably good way to view 2025’s capex. This implied Amazon is eyeing capex of ~$105 billion, up ~27% YoY after rising ~57% YoY in 2024. This would also be practically double Amazon’s $52.8 billion in capex from 2023, and comes after Amazon announced in early 2023 that it would be putting $150 billion towards new data centers over the next 15 years.  

Amazon Quickly Outpacing Hyperscaler Peers When it Comes to Capex  

We’ve closely tracked Big Tech’s capex for many quarters as a growth indicator for Nvidia and AI hardware suppliers. Amazon has been the largest spender of the hyperscalers (Amazon, Microsoft, Meta, Alphabet) over the last few years, accounting for nearly one-third of total capex in 2023 and 2024 and is set to be the top spender again in 2025.  

On a quarterly basis, Amazon’s capex accelerated significantly beginning in Q2 2024, rising above $15 billion for the first time before ending the year $10 billion higher. Amazon far outspent Google and Meta, though Microsoft was a heavy spender as well in the second half of 2024. 

Amazon is accelerating its capex as it continues to witness elevated AI demand that outstrips its available capacity. Jassy said both in Q3 2024, and in Amazon’s shareholder letter published in April 2025, that there was more demand AWS could fulfill even if they had more capacity, and that the “faster demand grows, the more datacenters, chips, and hardware we need to procure.” This suggests that AWS is continuing to see high levels of demand to give it the confidence to commit to $105B+ in capex for 2025. 

AWS is monetizing this capex “many months” after it is spent and “over many years,” suggesting that 2024’s acceleration to 19% growth is likely from monetizing investments from 2023 with Nvidia’s Hopper GPU generation. This also suggests that the surge in capex in 2H 2024 has not been monetized yet and AWS will recognize the growth benefits through 2025, especially as component constraints ease in the second half of the year.  

Data from Omdia shows Amazon was a significant deployer of both Nvidia’s GPUs and custom silicon in 2024, purchasing ~196K Hopper GPUs and deploying a combined 1.3 million Trainium and Inferentia chips (comparable to ~430K Hoppers). For 2025, Nvidia showed that the top CSPs (Amazon, Microsoft, Alphabet, Oracle) have already ordered 3.6 million Blackwell GPUs, versus 1.3 million at Hopper’s peak, with Amazon likely a large purchaser of the new generation.  

Based on management’s comments about the monetization timeline, it’s possible that Amazon is only just beginning to recognize the growth tailwinds from these large investments and GPU deployments. This also sets the stage for consistent, strong growth through 2026 as Blackwell comes online along with Trainium3.  

Despite this, there are still risks on the horizon –the recent tariff-fueled market turbulence has led to some heightened fears about the state of cloud spending, given that some customers may undergo budget optimization efforts in an effort to absorb higher costs in the coming months. For example, Mizuho’s James Lee predicts that up to 50% of cloud customers may eye budget reductions this year and be more “hesitant while assessing the economic impact before redeploying their capital.” 

A Note on Amazon’s AI Pricing, 3P Model Support versus Azure, GCP 

AWS is working to differentiate itself from Azure and Google Cloud (GCP) in two ways – offering a wide range of AI model support and on price-performance.  

AWS offers access to more than 100 leading foundation models, from DeepSeek, Mistral, Meta, and others, as well as access to Anthropic’s models, Amazon’s internal models, and bring-your-own model support in Bedrock and SageMaker. Azure is tightly coupled with OpenAI due to Microsoft’s partnership, though it does also offer access to leading open-source models such as those from DeepSeek and Meta; however, not all of OpenAI’s leading models are accessible in every region. Google offers access to its full suite of models with broad open-source support, though usage of Google models is locked to GCP. 

Amazon offers this range of model support to its entire custom silicon family of chips and Nvidia GPUs, while Azure lets customers build primarily on Nvidia’s ecosystem as its custom silicon effort is not nearly as extensive as AWS’. GCP mirrors Amazon with flexibility between TPUs and Nvidia GPUs.  

For chip pricing, AWS and GCP can both offer custom silicon instances for fraction of the cost of Nvidia GPU instances, while also offering cheaper prices for inference workloads. A report from CloudExpat estimated that Amazon’s Trn1 instances cost ~$1.34 per hour versus $1.20/hour for GCP’s TPU v5e and $12.84/hr for an H100 on Azure; for inference, an Inferentia instance was estimated at $0.40 per 1M tokens, versus $0.30 on GCP and $1.00 for the H100 on Azure.  

When it comes to running hosted inference on a model, such as for Meta’s Llama 405B, where costs range from as low as $6 per 1M tokens on Fireworks.ai to ~$21.33 on both Azure and AWS, all the way to $30 on Snowflake, Amazon’s performance sets it apart. While smaller platforms like Fireworks may hit users with ‘spike arrests’ where usage is throttled, Amazon can offer compellingly fast performance and token generation due to its breadth of instances. So for a model like Llama 3.1 405B, AWS says that independent inference performance tests “showed that Amazon Bedrock, running on Trn2 instances, delivers more than 3x higher token-generation throughput compared to other available offerings by major cloud providers.”  

Amazon’s $8B Anthropic Partnership, Project Rainier

Amazon is extending its competition with Microsoft to the AI startup sphere, backing OpenAI competitor Anthropic with a total investment so far of $8 billion in the Claude developer.  

Anthropic is quickly adding new models with improved capabilities and new features as it targets both consumer and enterprise AI demand as competition heats up. The startup unveiled an AI Research feature, which conducts multi-step autonomous research, with in-line citations and an ability to “synthesize findings and deliver holistic, source-backed summaries.” It also launched deeper integrations with Google Workspace, where Claude can now connect directly to Gmail, Google Calendar, and Google Docs, allowing it to scan emails, summarize docs, identify meetings, and surface files requested. 

Anthropic also recently introduced new subscription plans for Claude, building on top of its $20/month Claude Pro plan and supplementing its free tier. The new Max plan for $100/month offers 5x the usage as Pro, while the $200/month option offers up to 20x the usage. Additionally, Anthropic is said to be preparing to roll out a new “voice mode” AI assistant that users can speak to, as early as this month, following ChatGPT’s footsteps. The company is also working on a new major ‘hybrid’ model that can switch between deep reasoning and quick responses, offering developers a sliding scale customization tool to shift speeds and help reduce costs.  

Amazon is working to support Anthropic’s future growth with Project Rainier, a massive supercomputer being developed primarily to serve Anthropic’s growing compute needs. AWS is also optimizing its platform to offer faster access to Claude, with Bedrock offering a new "latency-optimized mode" for Claude 3.5 Haiku which runs 60% faster on Trainium2 instances. 

Project Rainier, in a way, is Amazon’s internal competitor to OpenAI and Oracle’s Project Stargate, as the supercomputer cluster, uniquely housed in multiple facilities in different locations linked by Amazon’s Elastic Fabric Adapter, will feature hundreds of thousands of its Trainium2 chips. Rainier is also unique in that it will be built entirely with Amazon’s custom chip stack, while heavily leveraging AWS’ Nitro system and its Neuron SDK for maximum performance. 

Rainier will be composed thousands of UltraServers linked together to form an UltraCluster accessing those chips. UltraClusters will help bring training times from weeks down to just a few days, according to AWS, as customers can access thousands of chips in tandem for large-scale training tasks.  

When completed, it will be one of the largest clusters ever built globally, though specifics about the exact amount of chips, cost and scale are still secret. Amazon is no stranger to large-scale cluster development though, having collaborated with Nvidia on Project Ceiba, a 20,736-Blackwell GPU cluster offering 414 exaflops performance.  

Investing in Anthropic can be seen as a net positive for Amazon despite the large capital outlays in its upfront investment and in Rainier – not only does Amazon have a guaranteed, leading AI customer to utilize the massive compute cluster it’s building, but it’s also receiving “intense feedback” from the startup on its custom accelerators and how to better optimize each new generation for performance or cost. 

AWS Margins Very Strong, but Risks Ahead 

AWS has historically enjoyed a strong operating margin profile that has strengthened considerably since early 2023, nearing the 40% range.  

In Q4, AWS reported an operating margin of 36.9%, down from its peak of 38.1% in Q3. This was a substantial improvement from a 29.6% margin in Q4 2023. Management noted that the strength of margins was due to strong growth and cost control – operating income for AWS has risen >30% YoY in each of the last six quarters, and 48% YoY in Q4. However, quarterly fluctuation is expected depending on the level of investment Amazon is making in the segment.  

On a TTM basis, AWS had only briefly surpassed 30% in the beginning of 2022 prior to a macro-induced growth slowdown that dented margins. Since then, AWS has managed to substantially improve its profitability, with TTM operating margin at 37% in Q4, up more than 12 points from 24.7% in Q2 2023.  

Should AWS continue to see operating income grow in the double-digits YoY, it could possibly reach the 40% threshold by the end of the year, providing a strong tailwind to EPS growth as it contributed half of Amazon’s total operating income in Q4.  

AI, Useful Life Classifications May Present Margin Headwind 

Management has talked about the AI versus non-AI margins, with AI margins being significantly lower due to the massive investments Amazon is undertaking at the moment. CFO Brian Olsavsky explained that AI “does come originally with lower margins and a heavy investment load,” and in the short-term that will be a headwind to margins, but over the long-term, he expects “margins will be comparable in non-AI business as well.” 

This is likely due to a much lower revenue return per dollar of spending presently, given that AI is only at a multi-billion dollar run rate. TD Cowen analysts put this in perspective, estimating that AWS historically has generated $4 in incremental revenue for every $1 of capital spending, but with surging AI investments, the ratio is now likely ~$0.20 cents for every $1. TD Cowen expects the incremental revenue to reapproach its usual $4 in the next several years. 

Another risk to operating margins that lies ahead is useful life calculations for servers, primarily that useful lives may continue to decrease as the pace of GPU upgrades accelerates, as the performance gaps between each generation widens.  

Amazon noted that AWS’ operating margin in Q4 benefited from a ~200 bp YoY positive impact from increasing the estimated useful life of servers in 2024. Stripping out this impact, operating margin would’ve been 34.9%. 

However, Amazon said that in Q4, they “completed a useful life study for our servers and networking equipment and observed an increased pace of technology development,” and as a result decreased the useful life for a “subset” of servers and networking from 6 years to 5 years. Management estimated that this change would decrease FY25 operating income by ~$700 million. 

Amazon also retired a subset of servers and networking equipment early, recording a $920 million accelerated depreciation expense, which they estimate will also negatively impact FY25 operating income by ~$600 million. Both of the impacts are expected to primarily be felt in AWS. This combines for a $1.3 billion negative headwind to operating income, which would be about a (2.7%) impact assuming 20% YoY growth in operating income. 

However, Nvidia is pushing ahead with a break-neck product upgrade cycle for its GPUs, maintaining an annual cadence that rival AMD is working to match. Each generation upgrade, such as from its Hopper generation to its Blackwell generation, promises significant leaps in compute and thus performance.  

These rapid performance upgrades could mean that current depreciation schedules still do not account for how quickly older generations of GPUs become obsolete simply from an inability to remain competitive performance-wise, whether that be in sheer compute or performance-per-watt. Should depreciation schedules more accurately be in the 3 to 4 year range, recognized depreciation expenses would be much larger, providing another headwind to operating margins.  

E-Commerce in Tariff Territory 

On the e-commerce side, Amazon is facing risks from increased tariff uncertainty, considering its rather high exposure to China. Online stores revenue growth has leveled off in the high-single digits, while third-party (3P) seller services revenue has decelerated rather sharply over the last few quarters. In light of the rising China risk, Amazon has also pushed forward with a large US warehouse expansion plan. 

According to Bloomberg, Amazon canceled several inventory orders from China after tariffs on the country were raised to 104%. The orders were said to contain numerous consumer-oriented products, such as air conditioners and scooters. Additionally, Amazon has high China exposure from first-party vendors, which account for ~40% of its items; Morgan Stanley estimates that 25% of the cost of goods sold from 1P vendors comes from China. There are also reports that Chinese vendors are considering hiking prices or leaving the marketplace due to the high tariffs.  

Tariffs also could weigh on 3P vendors, as higher import prices could squeeze margins and again lead to price hikes to offset higher costs. Amazon is said to have offered price concessions to some vendors with high-demand products in order to mitigate potential impacts.  

In light of the heightened tariff risk, Amazon is doubling down on its domestic US business, as it is reportedly looking for capital partners for a $15 billion project to expand its warehouse footprint. The plan would call for construction of 80 new logistics facilities, with the majority being delivery hubs and a few being multi-story, larger-scale fulfilment centers.  

  • North America revenue did reaccelerate to 10% YoY in Q4 to $115.6 billion, though growth has slowed over the past few quarters – Q2 and Q3 both saw the segment grow just 9% YoY, versus 11% YoY in the same period in 2023.  
  • International revenue decelerated rather sharply in Q4, with growth of 9% YoY to $43.4 billion, down from 12% growth in Q3. This would mark the segment’s slowest growth since the start of 2023.

Segment Breakdown: 

Decelerations in 3P seller services (which includes commissions, fulfillment and shipping fees) and advertising occurring simultaneously raise red flags for Amazon heading into Q1, as it suggests that its seller flywheel may be losing some momentum as tariff risks escalate. The softness of 3P seller services noted below hints that marketplace volume may be slowing, while the slowdown in advertising growth (in the holiday quarter) could mean sellers are spending less on advertising due to weaker sales performance, tighter margins or lower product demand. 

  • Amazon’s online store revenue was $75.6 billion in Q4, up 8% YoY, maintaining the same growth pace from Q3 yet accelerating from the 6% growth seen in Q1.  
  • On the other hand, 3P Seller Services growth has decelerated dramatically, from 19% in Q4 2023 to just 9% in Q4 2024.  
  • Advertising growth has mirrored that deceleration, from 26% in Q4 2023 to just 18% in Q4 2024. 

Recent Strengthening in E-commerce Margins Provides Some Cushion for Tariffs 

Tariff risk looks to be amplifying some underlying weaknesses in 3P seller services and advertising; two signals of marketplace volume and demand. However, Amazon does have levers it can pull to mitigate these effects, as it is recognizing cost savings in inventory management and has a much stronger operating margin profile now that can absorb some costs. 

Amazon’s management explained in Q4 that some genAI applications they built for inventory management have led to “10% better forecasting on our part and 20% better regional prediction,” and improvements in robotic fulfilment efficiency have combined for “significant productivity and cost savings.” Management sees more opportunities to further reduce costs via more refined inventory placement, expansion of same-day delivery networks, and accelerated robotics and automation throughout the fulfillment process.  

On the margin side, Amazon has seen solid improvement in both North America and International that should offer it some leeway when it comes to absorbing tariff impacts: 

  • North America operating margin reached 8.0% in Q4, up nearly 2 points YoY, while TTM operating margin reached 6.4%, up 2.2 points YoY.  
  • In dollar terms, North America’s TTM operating income rose 68% to $25.0 billion, or a $10.1 billion YoY gain.  

International operating margin was 3.0% in Q4, a 4 point YoY improvement from (1.0%), while TTM operating margin was 2.7%, a notable improvement from (2.7%) last year. 

Revenue Guide Misses Estimates 

Amazon’s soft Q1 guidance pointed to growth in the high-single digit range, a sharp sequential deceleration and what would be the first single-digit print in the last eight quarters.  

For Q1, Amazon projected revenue between $151 billion and $155.5 billion, for 5% to 9% YoY growth; which was well below the consensus estimate for $158 billion.  

Amazon said the softness was partially due to an “unusually large, unfavorable impact of approximately $2.1 billion” from foreign exchange rates (150 bp YoY growth), as well as the lack of leap-year impact, which added $1.5 billion of revenue (or 120 bp YoY growth impact). Even backing out both headwinds, growth would still be projected in the single-digits using the midpoint of the range. It also likely does not account for increased macro turbulence on the retail side as the full breadth and severity of tariffs on key trading partners was not known when guidance was provided. 

At the midpoint of 7% YoY growth, this would be Amazon’s slowest quarterly growth rate since Q3 2001. This also does not account for some of the escalated tariff risks and seller impacts that have arisen throughout April. Analysts expect Amazon to fare much better, with the current consensus estimate calling for 8.2% growth to $155 billion in revenue, at the upper end of the guided range. Analyst estimates range from $153.2 billion to $157.2 billion, with none of the 45 analysts expecting Amazon to report below 7% growth this quarter.  

EPS Growth Outpacing Revenue on Margin Strengths  

Despite the weak revenue guide and sequential deceleration, Amazon’s margin strengths, primarily from AWS but also from improvements on the e-commerce side, are aiding robust earnings growth.  

Amazon is currently estimated to generate $1.38 in EPS in Q1, up nearly 39% YoY, or more than 5x the rate of revenue growth at midpoint. EPS has rebounded from 2023’s lows, rising 216% in Q1 2024 and >50% YoY in each quarter of 2024.  

Q1’s EPS estimate does face some risks from tariffs impacting margins or weaker revenue growth, while broader macro risks in consumer spending are becoming more prevalent. Consumer spending fell by the most in four years in January, and while it rebounded in February, much of the rebound was driven by price increases as inflation-adjusted spend barely rose. Consumer spending is being closely watched as consumers remain a bit more cautious, with Richmond Fed President Richard Barkin saying that while spending is not yet showing “troubling signs of a decline,” it is the metric he is most closely watching as the “trigger” on the economy.  

EPS growth is forecast to slow rather dramatically as Amazon continues to face tough comps and slower growth – each quarter in 2025 is expected to see sub-10% revenue growth. Q2 and Q3 are expected to see EPS growth of just 12% to 13%, while Q4 runs the risk of falling flat, with estimates pointing to just 2% YoY growth. 

For 2025, EPS growth is projected to outpace revenue growth by 5 points at just over 14% YoY, as Amazon benefits from the improved margin base it built throughout 2024, which drove EPS growth of 91% YoY last year.  

Gross margin is on the verge of breaking 20% as high-margin AWS and advertising increase their revenue share, while operating margin sharply expanded to the double-digits for the first time, up from the low single digits at the start of 2023. Should Amazon be able to drive and maintain an operating margin >40% for AWS, it is well on the path to see mid-20% gross margins and mid-teens operating margins over the next few years.  

While Amazon is lapping strong growth this year, EPS growth is expected to accelerate through FY27 on the back of this margin strengthening. FY26 EPS growth is currently projected at 19% YoY before accelerating to nearly 25% YoY in FY27, representing a 10 point acceleration in two years. 

From FY25 to FY27, EPS growth is currently around a 19% CAGR. With the way that AWS is quickly reshaping Amazon’s margin profile and putting it on a trajectory for mid-teens operating margins, Amazon could likely see EPS growth accelerate to low to mid-20% CAGR simply from the growing margin and profit tailwinds from AWS.   

Operating Cash Flows Remain Robust, FCF Impacted by Capex 

Amazon has driven significant growth in operating cash flow, though FCF growth has been minimal due to accelerating capex. 

Amazon generated $115.9 billion in operating cash flow in 2024, up 36% YoY. OCF margin expanded 3.4 points to 18.2%. This has been a remarkable turnaround from 2022, where TTM OCF dropped to just $35.6 billion.  

However, FCF growth has stalled, with Amazon reporting minimal FCF generation in the first three quarters of 2024 due to sharply increased capex. For 2024, FCF rose just 4% YoY to $38.2 billion, but on a TTM basis, FCF has declined sharply from its peak at $53 billion in Q2. This was because Amazon reported just $4.7 billion in OCF in Q3 and $5.1 billion in Q1, or margins of just 3% and 3.5% respectively.  

While FCF’s growth trajectory had tracked OCF rather closely through 2021 all the way to 2024, FCF has detached and began declining due to capex acceleration. With management signaling ~$26 billion in capex each quarter in 2025, it’s likely that FCF generation and growth will lag OCF through next year.  

Valuation at Historic Lows 

Amazon is trading at historically low valuations on an earnings and cash flow basis, while it continues to benefit from improving operating leverage as AWS and ads help drive margins and earnings growth higher.  

Amazon is trading at a <29x forward earnings, at 2024 and below 2022 lows, which would be the lowest forward P/E ratio the company has traded at in more than a decade. It’s also well below the 36-38x multiple it commanded through much of 2024, and a ~30% discount to its 5-year average forward P/E of 39.4x.  

Given the strong growth in OCF, it is no surprise that Amazon is trading at its lowest P/OCF multiple at just 17x. This is also a 30% discount to its 5-year average P/OCF multiple of 25.7x, and a 20% discount to retail peer Walmart, which is trading at 20.6x OCF for just 2% YoY growth compared to Amazon’s 36%.  

While Amazon is trading at historic lows for forward P/E and P/OCF, its top-line multiples have been consistently expanding as high-growth, high-margin AWS and advertising continued to grow their share of revenue and boost Amazon’s earnings power. Amazon is currently trading at 2.8x forward P/S, nearly double its multiple from early 2023 but in line with its 5-year average at 2.7x.

Conclusion 

E-commerce made Amazon famous, yet its AWS segment is quietly and quickly becoming a driver of Amazon’s growth story.  

Amazon is positioning AWS to become an AI behemoth with a focus on offering everything customers will need to fully harness AI efficiently and cheaply. AWS continues to launch new Trainium and Inferentia instances to offer a broader range of powerful, affordable training and inference-optimized compute resources to its customers. It also provides access to 100+ leading models and bring-your-own model support, while launching 2x more AI features than other CSPs combined from early 2023 through Q3 2024. Project Rainier is showcasing AWS’ powerful in-house chips and networking, aiming to scale up to hundreds of thousands of chips housed in different facilities as one of the largest GPU clusters built to date. 

Capex spending hints that growth for AWS could remain strong through 2025 and into 2026 as Amazon begins to monetize 2024’s accelerating investments. While the e-commerce side bears the brunt of potential tariff impacts, Amazon’s valuation has gotten attractive recently as P/E and P/OCF multiples have fallen toward historic lows. 

Want exclusive, fast-paced deep-dive research on 30-40 stocks per year in the market’s hottest sector, from AI to data center, software, crypto and more, from a leading tech portfolio? This is what you’ll get with the I/O Fund’s new Discovery tier.  

Sample research we published in March and April:   

  • Key supplier to TSMC’s new high-growth platform called Compact Universal Photonic Engine (COUPE)  
  • Nuclear and natural gas supplier for AI data centers   
  • A breakdown of the risks and opportunities for the biggest IPO in the AI sector

Current Pro and Advanced Members: To subscribe to Discovery with 40% off, please click here to email usclick here to email us or email premium@io-fund.com and mention code DISCOVERY40.

Please note: The I/O Fund conducts research and draws conclusions for the Fund’s positions. We then share that information with our readers. This is not a guarantee of a stock’s performance. Please consult your personal financial advisor before buying any stock in the companies mentioned in this analysis.

Recommended Reading:

Tesla Stock Faces Recalibration of Growth Expectations

Posted on by io-fund

It was no surprise that Tesla’s Q1 deliveries came in below estimates, given the multitude of data points across Europe and China that pointed to significant YoY declines in demand. We pointed out in the first week of March when we last covered Tesla’s stock that given this weakness in core regions and transitory production impacts from the refreshed Model Y, there was a risk that Q1 deliveries would fall to well below 400,000 vehicles, whereas consensus estimates at the time were calling for more than 410,000.

Tesla’s stock is now facing a recalibration of expectations after Q1’s delivery report missed by a wide margin, with revenue and EPS growth estimates falling sharply into its earnings report next week. Below, I dig into the risks Tesla’s revenue faces in Q1, growth expectations for the year, and the bigger picture ahead.

Tesla’s Q1 Deliveries Miss Mark

Tesla reported 336,681 deliveries in the first quarter, more than 40,000 shy of the 377,592 estimate, representing the worst performance in two years. This represented a (13%) YoY decline in deliveries, while production fell (16%) YoY to 362,615 vehicles, the lowest quarterly production since Q2 2022. Tesla said that the transition to the refreshed Model Y led to “several weeks” of lost production.

Graph of Tesla stock quarterly production and deliveries showing (-16%) YoY decline in production and (-13%) YoY decline in deliveries/

Q1 saw a sharp decline in Tesla’s production and deliveries to the lowest levels since 2022. Source: I/O Fund

These delivery and production figures provide a few clues for the upcoming earnings report and for Q2 as well:

1) Outsized revenue impact due to ASP deterioration and delivery decline

2) Margin headwinds, not just from idle capacity and the refreshed Model Y ramp, but also incentives

3) Possible inventory build-up with production outpacing deliveries by ~26K vehicles, which may force more aggressive financing activity to sell down excess inventory in Q2

4) Recalibration of expectations for 2025 growth

Recalibration of Growth Expectations

The delivery weakness looks to be snowballing into something much larger – a recalibration of growth expectations for Tesla for the full year. Even in the six weeks from our previous update, revenue and EPS estimates for Q1 and for 2025 have continued to drop considerably, marking a rapid shift in expectations impacted by the weak delivery print.

At the start of 2025, Tesla was expected to report $25.98 billion in revenue in Q1, up 22% YoY, with EPS of $0.71, for YoY growth of 58%. These estimates had been little changed since July 2024.

Now, less than a week from Tesla’s report, Q1 revenue is pegged at $21.54 billion, for growth of just 1.1% YoY. Q1’s EPS estimate is $0.43 for a (4%) decline.

Graph showing Tesla stock's Q1 2025 revenue and EPS estimates falling significantly since the start of the year.

Tesla’s Q1 revenue and EPS estimates saw sharp changes heading into the end of the quarter, with revenue estimates coming down more than $2 billion in March. Source: YChartsYCharts

Essentially, in just four months, Tesla has shaved off 20 points of revenue growth and more than 60 points of EPS growth, assuming it meets estimates for Q1. This is a sudden, sharp recalibration of growth expectations for the quarter that has carry-over effects for full-year growth.

Now to the full year: at the start of 2025, Tesla was estimated to generate $116.7 billion in revenue, up nearly 19.5% YoY, with EPS of $3.15, up 30% YoY.

Now, 2025 revenue is expected to be $106.7 billion, $9 billion below where it entered the year and representing growth of just 9.4% YoY, while EPS is estimated to be $2.55, (19%) lower than at the start of the year and representing growth of just 4.6% YoY.

Graph showing Tesla stock's 2025 revenue and EPS estimates falling significantly since the start of the year.

Tesla’s revenue and EPS estimates have fallen sharply since the beginning of the year, with EPS estimates now 19% lower and revenue now nearly 8% lower. Source: YChartsYCharts

In just four months, Tesla’s full year revenue growth estimate has fallen 10 points to the high-single digit range, while its EPS growth estimate has declined more than 25 points to the mid-single digit range. This is not solely due to Q1, as Q2 and Q3 estimates have been revised (2%) to (4%) lower as well after Q1’s weak report.

ASP Weakness Exacerbating Tesla’s Delivery Decline

Q1’s revenue is likely to be pressured as weak pricing weighs on Q1’s weak deliveries, providing additional challenges for automotive revenue, its primary revenue segment at 77% of revenue in Q4.

Tesla’s automotive revenue faces amplifying headwinds in Q1, with weaker ASPs weighing on an already large (13%) YoY decline in deliveries. This presents broader growth challenges given that Energy Storage revenue has decoupled from deployment growth, and as a result, Tesla may run the risk of revenue coming in below $20 billion in Q1.

Here’s why:  

Tesla has not increased its ASPs on a QoQ basis since Q2 2022. Compared to last year, ASP was ~$3,700 lower YoY in Q4. This alone presents a 9-point additional headwind to automotive revenue growth (excl. leasing and reg credits) simply due to the swift pace of price declines.

In Q1, ASPs faced a complex environment, from continuing promotional efforts (0% APR loans, referral discounts, etc), combined with price hikes for the Model X, slight increases in average transaction prices in the US near the end of the quarter, and the launch of the Model Y Juniper, which may actually provide a drag to ASP this quarter.

Impact of Tesla’s New Model Y

Model Y Juniper’s popularity in China may dent ASP in Q1, as the new variant is priced lower than its US counterpart: the RWD variant is priced at 263,500 yuan, or ~$36,000, while the Long Range AWD is priced at 303,500 yuan, or $41,500, versus $48,990 for the LR AWD in the US before federal tax credits.

The new model reportedly saw nearly 60,000 orders in the first 5 days in China, becoming the country’s top-selling model in March with 43,370 deliveries. The RWD variant is priced nearly 10% below Tesla’s Q4 global ASP in US$ figures, while the AWD variant is just over 3% higher, suggesting that if a majority of sales volumes were for the cheaper RWD variant, global ASPs will face some pressure.

While Juniper carries a higher price tag than the previous Model Y variant, it was only available in the US in mid-March, leaving little time for deliveries to commence in volume. Tesla also has reportedly discontinued the Launch Series, meaning the new model no longer has its $11,000 more expensive option available for sale, limiting its ability to positively impact ASPs.

Tesla Could See Q1 Revenue Below $20 Billion

Assuming ASPs remain approximately flat QoQ due to the aforementioned factors – slight price hikes, base Juniper variant priced below global ASPs in China with strong deliveries, and financing/incentive offers – automotive revenue could decline nearly (22%) to $12.9 billion. Including approximately $1.2 billion in leasing and regulatory credits, total auto revenue could be just under $14.1 billion, or down (19%) YoY.

Energy Storage decoupled from deployment growth in Q4, as revenue grew less than 2% from Q2’s peak despite 17% higher deployments. On a YoY basis in Q4, deployments rose 244%, yet revenue rose just 113% YoY, a staggering 131 point difference pointing to pricing pressures. This suggests that growth is likely to remain pressured in Q1, as Energy deployments declined more than (5%) QoQ in Q1 to 10.4 GWh.

Should Energy Storage revenue follow that pattern of revenue growth weaker than deployment growth, at something such as a (8%) QoQ decline, or up 73% YoY, while Services revenue increases 4% QoQ, this would project total Q1 revenue out to $19.85 billion.

Even assuming 10% QoQ growth in Energy Storage revenue, or up 106% YoY, a 2% increase in ASP (this would be the first time in 11 quarters), $1.2 billion in leasing and regulatory credit revenue, and the 4% QoQ Services growth, revenue would project out to $20.6 billion.

To hit the current consensus of $21.54 billion, which has declined from $22.46 billion since deliveries were reported, Tesla would need a massive 7.5% QoQ increase in ASP, or up $3,000 sequentially, plus 10% QoQ revenue growth in both Energy Storage and Services, with $1.2 billion in reg. credit and leasing revenue, an unlikely scenario.

The I/O Fund specializes in covering lesser-known AI stocks on our research site with trade alerts and weekly webinars. Learn more here.The I/O Fund specializes in covering lesser-known AI stocks on our research site with trade alerts and weekly webinars. Learn more here.

Lingering Margin Pressures – New Lows Likely Ahead

Tesla is facing a slew of headwinds to margins, as production and delivery numbers are suggesting inventory buildups, alongside known impacts from ramp-related costs. The combination of weaker deliveries and weaker automotive revenue growth is also likely to weigh on already thin margins.

Q4 already saw increased margin pressures from aggressive financing efforts to clear out excess inventory, and average gross profit per vehicle was on the brink of falling below $5,000 as a result. In Q1, Tesla now must contend with much weaker automotive revenue on top of additional ramp related costs for the refreshed Model Y. Ramping Megapack and Powerwall output in Shanghai also is likely to weigh on operating margin in the quarter.

For Q2, margins are likely to face additional pressure from tariff impacts. Tesla acknowledged that tariffs “will have an impact on our business and profitability,” and while the administration has floated “possible temporary exemptions” on imported auto parts and vehicles, nothing is set in stone. Reports from Anderson Economic Group estimate that for US-assembled small crossovers, sedans, and mid-sized SUVs, tariffs costs could be $2,500 to $4,500.

According to NHTSA, for the 2025 model year, Tesla’s vehicles source 20% to 25% of its part content outside the US, primarily from Mexico, opening it up to some tariff risks. Additionally, Tesla was said to have suspended plans to source some components for the Cybercab and Semi following increasing China tariffs, disrupting production timelines for the two vehicles, which were originally expected to start trial production in October.

1.48M More Deliveries Needed to Return to Growth in 2025

We said in our March 6 analysis, Tesla Has a Demand Problem; The Stock is Dropping, that if Q1 did come in weak due to global sales weakness and transitory production impacts, Tesla would have to “make up substantial ground in the back half of the year” to reach optimistic delivery targets.

As you may recall, Tesla quietly shifted its tone on 2025 deliveries growth, with Musk originally stating that 20% to 30% growth was achievable in 2025 in Q3 2024’s earnings call, while Q4’s earnings call shifted the outlook to just a ‘return to growth’.

Now that Q1 deliveries have been officially announced, we can see where Tesla stands on its path to growth. 2024 deliveries totaled 1.81 million, meaning Tesla would need to deliver 1.48 million more vehicles from Q2 to Q4 to return to growth this year.

Q2 is unlikely to ease growth fears, as expectations similarly have been lowered quickly as uncertainty around tariffs and vehicle demand looms. Initial estimates currently point to deliveries between 350,000 to 375,000 for Q2, versus nearly 444,000 in the year ago quarter, though it is still quite early in the quarter and subject to change.

Tesla reportedly halted shipments of Model X and S vehicles to China due to high tariffs, though this is expected to have a minimal impact on deliveries considering the X/S accounted for <5% of global deliveries last year, with China a portion of that. As a whole, China is a major market for Tesla accounting for approximately 37% of deliveries last year, and rising geopolitical tensions could put some of that at risk.

Returning to growth in deliveries this year would require Tesla to deliver over 491,000 vehicles on average each quarter for the rest of the year, based on how Q1 started. This would require Tesla to increase deliveries at ~20% QoQ each quarter, which would culminate in 580,000 in Q4, or a fresh record and pointing to 17% YoY growth, a figure that Tesla has not hit in any of the past nine quarters. Put another way, given Tesla’s annual capacity of >2.35 million vehicles, or 587,000 quarterly, Tesla would need to be producing and delivering at almost maximum capacity come Q4.

The Bigger Picture Ahead for Tesla

As we said in our March analysis, price often bottoms before fundamentals, and it’s clear that Q1 and even Q2 are expected to be the weakest point of the year before the fundamentals improve heading into 2026.

As it stands currently, Tesla is expected to see revenue growth significantly reaccelerate beginning in the back half of the year, with EPS following in 2026. Estimates point to growth bottoming at 0.3% in Q2, before accelerating to 14.3% YoY to end 2025 and 25.7% in Q1 2026, representing a quick 25 point growth acceleration in four quarters.

Graph of Tesla stock quarterly revenue and EPS growth estimates through Q1 2026, showing revenue accelerating from 1% YoY to 26% YoY.

Tesla’s revenue growth is currently forecast to bottom in Q2 before accelerating to nearly 26% in Q1 2026. Source: I/O Fund

EPS growth is expected to be choppy this year, with only Q2 expected to see growth against a much weaker comp. However, Q1 2026 is expected to see the sharpest YoY growth in EPS in more than three years at 54.2% — again against a soft comp.

We have continuously reminded our readers that margins have been the #1 metric to focus on for nearly two years as margin compression led to 2024’s large YoY declines in EPS. Tesla has had to continue cutting ASPs to promote vehicle affordability, not relieving any of this margin pressure.

In Tesla’s case, however, sentiment often takes precedence above all, in part due to a massive advantage it has in physical AI and (speculative) trillion-dollar opportunities in robotaxis and humanoid robotics with Optimus.

Tesla is witnessing exponential growth in FSD-driven miles, nearly tripling from ~1 billion in March 2024 to almost 3 billion by December 2024. It is also still planning to launch a driverless robotaxi service this year and begin scaling Optimus production. While neither of the two are expected to be meaningful contributors to growth for this year, the materialization of either AVs or humanoid robotics could open tremendous growth potential for the company in the future.

Timing has always been tricky for Tesla, especially when it comes to autonomous vehicles. However, timing investments is one of the I/O Fund’s strengths, actively managing positions with real-time trade alerts for every buy and sell we make in our leading portfolio with 210% returns since inception and 27.6% annualized returns. We also offer frequent deep dive research, weekly webinars and an automated portfolio hedge. Learn more here.

I/O Fund Equity Analyst Damien Robbins contributed to this report.

Please note: The I/O Fund conducts research and draws conclusions for the Fund’s positions. We then share that information with our readers. This is not a guarantee of a stock’s performance. Please consult your personal financial advisor before buying any stock in the companies mentioned in this analysis.

Recommended Reading:

Potential Upstream Beneficiary of Nvidia’s CPO Push Come 2026

Posted on by io-fund

There is a small-cap company that is emerging as a potential beneficiary of an upcoming shift to co-packaged optics over the next two years. Nvidia introduced its first co-packaged optics (CPO) switches with integrated silicon photonics at GTC 2025, addressing the need of providing more bandwidth and faster speeds while reducing energy consumption in exascale GPU clusters.  

Nvidia is utilizing TSMC’s COUPE (Compact Universal Photonic Engine) for its Quantum-X and Spectrum-X switches, and this company’s wafer-level optics tech (WLO) is expected to be an integral part of the CPO supply chain for TSMC’s COUPE.  

For 2025, this company is expecting little to no revenue from CPO, though it expects the business to ramp quite rapidly come 2026 and beyond, due to the advantages CPO offers in terms of bandwidth and its upstream placement in the supply chain serving a critical need for COUPE.  

While this analysis will focus on this stock’s opportunities in wafer-level optics, which has the potential to quickly become a significant driver for revenue growth from the emergence of co-packaged optics, it’s important to note that this segment remains quite small presently while automotive remains the core growth driver for the company.  

Every Thursday at 4:30 pm Eastern, the I/O Fund team holds a webinar for premium members to discuss how to navigate the broad market, as well as various stock and crypto entries and exits. Beth Kindig offers weekly deep dives including lesser-known cryptocurrencies and AI stocks, plus the team offers trade alerts. 📈 The I/O Fund team is one of the only audited portfolios available to individual investors. 📊 If you’d like to subscribe to the Advanced Market Signals plan, email us at premium@io-fund.com. 📧4:30 pm Eastern, the I/O Fund team holds a webinar for premium members to discuss how to navigate the broad market, as well as various stock and crypto entries and exits. Beth Kindig offers weekly deep dives including lesser-known cryptocurrencies and AI stocks, plus the team offers trade alerts. 📈 The I/O Fund team is one of the only audited portfolios available to individual investors. 📊 If you’d like to subscribe to the Advanced Market Signals plan, email us at premium@io-fund.com. emium@io-fund.com. 📧

Nvidia Unveils New CPO Switches at GTC 

The AI industry has been rapidly progressing towards million-plus GPU clusters, with Broadcom outlining late last year that it believes its three hyperscale customers are aiming to each have 1-million GPU clusters by 2027, which will require new scale-up networking architectures and components. 

Co-packaged optics (CPOs) are seen as one possible way to break past the cost and power bottlenecks of current pluggable optical technologies and facilitate the development of larger-scale AI clusters. CPOs place optics directly on the switch die which decreases the distance and thus insertion loss, while improving power efficiency due to that shorter distance. 

Source: Cisco 

Nvidia’s two new switches unveiled at GTC, the Quantum-X Photonics InfiniBand and Spectrum-X Photonics Ethernet, were designed with that million-plus GPU cluster scale goal in mind. Nvidia says that by replacing pluggable optics with silicon photonics on the package, it can “deliver 3.5x more power efficiency, 63x greater signal integrity, 10x better network resiliency at scale and 1.3x faster deployment compared with traditional methods.” In simple terms, the switches are more reliable, more efficient, consume less power, and are quicker to deploy than current networking tech. 

Nvidia VP Ian Buck stated at GTC that the CPO switches help reduce power consumption by eliminating the need for external lasers and pluggable transceivers to achieve a significant reduction in power from 39 watts to 9 watts. Buck explained that this “gives you that benefit from going from 39 watts of power down to only 9 watts of power for the same number of ports, and that's huge. It doesn't sound like 39 sounds a lot. But if you get 400,000 GPUs in an AI supercomputer, there's like 24 megawatts of lasers like so that's a lot of laser light that could be optimized and made more efficient.” 

24 MW may not sound like much, but given that the GB300’s are expected to draw up to 140 kW per rack, that’s the equivalent of nearly 172 GB300 racks (or the equivalent of approx. 12,400 individual GPUs) in power savings from the shift to CPO. Thus, by shrinking the bill of materials by eliminating pluggable transceivers while simultaneously providing a much more power-efficient solution, Nvidia can further reduce TCO for its customers. 

Nvidia Expects CPO Switch Availability as Soon as 2H 2025 

Similar to its GPU roadmap, Nvidia is moving quickly when it comes to its CPO switches, expecting availability for its Quantum-X Photonics InfiniBand switches as soon as the second half of this year, while its Spectrum-X Photonics Ethernet switches are expected to be released in 2026.  

TSMC’s COUPE Paves Way for CPO Adoption 

Nvidia’s new switches rely on TSMC’s silicon photonics platform called COUPE (compact universal photonic engine).  

TSMC first published research on this new platform in 2021, as a team of researchers discussed a primary challenge of mass adoption of silicon photonics, which was at the time the lack of an integration platform that could meet a range of power, performance and cost needs. TSMC’s researchers proposed COUPE, which they said had the ability to meet “the most demanding system requirements and pave the way for [silicon photonics]-based wafer level system integration (WLSI) for high performance computing applications.”  

COUPE combines a photonic integrated circuit (PIC) with a 65nm electronic integrated circuit (EIC) via its SoIC-X advanced packaging tech. TSMC says that utilizing SoIC-X “enables the lowest impedance at the die-to-die interface and therefore the highest energy efficiency.” COUPE is also easily integrated with merchant ASICs and GPUs to form a co-packaged structure.  

TSMC laid out a tentative timeline at the 2024 North American Technology Symposium in April 2024. At the symposium, TSMC laid forth a plan to enable COUPE in pluggable optics in 2025, and COUPE on substrate in a CoWoS-based CPO in 2026, which it said would offer 2x reduction in power and 10x reduction in latency. TSMC mentioned that it was exploring COUPE on CoWoS interposers for further power reduction.  

Recently, TSMC has been rather tight-lipped about COUPE and co-packaged optics — in Q4’s earnings call, CEO C.C. Wei was questioned about CPO and how it would facilitate this shift in the supply chain. Wei answered that for “big volume, I don’t think it will be in this year, or probably we have to wait for one or one and a half year to see that contribution or the volume production. The initial results are quite good, no doubt about it.” 

SoIC capacity is expected to rise at a rapid rate, though there have not been any major updates recently. At the beginning of 2024, TSMC’s SoIC capacity was estimated to reach 5,000 to 6,000 wafers/month by the end of the year, a 150% to 200% YoY increase. 2025’s capacity at the time was expected to reach 10,000 wafers/month, nearly doubling YoY. A recent update from TrendForce reaffirmed the 10,000/month capacity for 2025 and outlined the possibility of another doubling of capacity in 2026. 

It had been reported in May 2024 that TSMC was aiming to expand its SoIC capacity at a 100% CAGR from 2023 through 2026, correlating to an “eight-fold” increase in capacity from 2023’s levels by 2026. 

FOCI’s Role in CPO with TSMC 

While TSMC is taking a more back-seat view to CPO’s timelines, suppliers are expecting a more accelerated ramp, with FOCI and Browave both seeing initial shipments in 2025 and a rapid ramp through 2026. FOCI is viewed as one of TSMC’s closest collaborators on CPO.  

In its 2023 annual report, FOCI stated that “CPO technology has just entered the market, and the production cost is still high,” but with the “explosion of high-speed transmission demand, it is expected that CPO technology will be a necessary technology that cannot be ignored and will enter the market in large quantities after 2025.”  

According to DigiTimes, FOCI anticipates that CPO fiber array products will see some smaller scale shipments in 2025, leveraging its ReLFACon (Reflowable Lensed Fiber Array Connector) products for optical switches and HPC/AI use cases. Mass production of CPO components is expected in 2026.  

FOCI is rumored to be the sole supplier of external fiber array units (FAUs) to TSMC’s first and second-gen COUPE via its ReLFACon product. FAUs are a critical part of CPO tech — the FAUs ensure smooth and efficient coupling of optical fibers to the silicon photonics engines, which enables high data throughput and reduces latency. In COUPE, the FAUs are used to align optical fibers on the PIC; COUPE 2.0 introduces a broad-band coupler that integrates with the FAUs to minimize insertion loss and extend alignment tolerance, which can lower manufacturing costs at the alignment stage. 

FOCI is not the only supplier preparing for strong growth come 2026 — Browave (which also holds a 4.6% stake in FOCI) is said to be fast-tracking its CPO development, as one of the named partners for Nvidia’s switches at GTC. Browave “expects to complete CPO validation by late 2025,” and is preparing for a “breakout” 2026, per DigiTimes.  

Explosive Growth Forecasted for CPO Market 

The CPO market is relatively new in that 2025 is expected to be the first year when growth surfaces before accelerating through 2026. Morgan Stanley estimated the CPO market size at just $8 million in 2023, projecting a 172% CAGR through 2030 in its base case scenario to reach $9.3 billion.  

This base case forecast has a few key assumptions: 1) Nvidia is the first to adopt CPO in its Rubin rack systems in 2026, with Rubin’s shipment volume reaching 200,000 units in 2026 and 700,000 in 2027; 2) other vendors including Cisco, Broadcom and Marvell begin shipments in 2027. 

Morgan Stanley also outlined a bull case scenario, projecting growth at a 210% CAGR, resulting in a significantly larger end market of $23 billion by 2030. This relies on much more optimistic assumptions: 1) Rubin shipments of 500,000 in 2026 and 1.75 million in 2027; 2) broader adoption of CPO by more chip manufacturers; 3) CPO yield rates greatly improved. 

However, Morgan Stanley’s bear case scenario sees the CPO market only $1.3 billion by 2030, on the assumptions that CPO yield issues cause shipment delays and thus a lower customer willingness to adopt the technology.

CPO One of 3 Growth Opportunities Ahead for Himax  

Himax is expected to be a critical player in this CPO push due to its partnership with FOCI, and sees CPO and WLO as one of three different growth opportunities ahead (with the other two being automotive OLED and ultralow power WiseEye AI sensing chips). For Himax, WLO presents a rather large opportunity over the next six to twelve quarters as Nvidia progresses with CPO switches to enable the scaling of data centers to millions of GPUs.  

In June 2024, Himax deepened its strategic partnership with FOCI and acquired a 5.3% equity stake in the firm for $16 million. Himax explained that the partnership leverages and combines WFO’s expertise with FOCI’s advanced ReLFACon solution to develop linear pluggable optics (LPO) and co-packaged optics solutions for AI and HPC chips that “demand enhanced bandwidth, improved data rate, minimized signal loss, reduced latency, and lower energy consumption.”  

Himax says that its WLO technology plays a crucial role in CPO by “providing essential optical coupling capability” as a core part of the CPO solution. Himax is said to be a sole supplier of micro-lens array units to FOCI’s FAUs for first and second-gen COUPE. 

Himax Outlines CPO’s Value Proposition, Ramp Profile 

Shortly after this investment, in August’s Q2 earnings call, management noted that they expect WLO to play an “even more decisive role in the next-generation optical technology landscape,” due to its versatility, precision and “small form factor characteristics that are not feasible with alternative technologies.” They stated that they believe LPO and CPO tech can “generate new, long-lasting revenue streams for Himax.” 

Management added that they are working closely with FOCI to align with multi-year roadmaps of foundry partners and AI chip customers with an effort to meet their near-term production targets. It is assumed that this comment is referring to TSMC’s COUPE and Nvidia’s CPO switches.  

 Q2’s analyst Q&A highlighted that this is not simply an R&D project, but rather one that is progressing quickly towards mass production: 

Q: Donnie Teng, Nomura 

My second question is regarding to the CPO. So would you maybe elaborate more on, you know, what’s the timeline of the CPO product. When should we expect to see some, you know, small volume contribution? And how confident you are to ramp up this business in the mid to long term?  

A: Jordan Wu, Himax CEO 

“First on precise timeline, I’m bound by NDA of my partner and customer, so I’m afraid I cannot give you very, very specific date or timetable. But I can tell you, what we are working on right now, the design is targeting for mass production. It is certainly not a R&D concept. …. Actually we’re way past the stage … and we are now pushing towards mass production ASAP. That’s what I can tell you. In fact, we expect to see some small but very early result hopefully by the end of this year but that’s minimal. But the next year, you know, if everything goes as planned, there will be steady ramping. And the confidence level mid to long term, I would say very confident. 

… So I would say, you know, everybody in the ecosystem is very keen to making sure that this happens ASAP. … And we have a roadmap together with partners, our customers, to really pretty dramatically expand the transmission bandwidth very substantially. I’m talking about by multiple times over the next few years. And you know, some of these projects are already in experimental stage in the earlier experimental stage or more mature experimental stage.” 

Himax confirmed in its 2024 annual report that small-scale production of its first-gen CPO solution was already underway by the end of 2024.  

Himax Provides Major Update in Q4, Lays Out CPO Revenue Opportunity 

Himax provided a major update in Q4 regarding CPO, as it continues to progress through small-scale production of its solution for LPOs and developing its platform for CPO architectures with FOCI. 

Management stated that the long-term prospect of CPO remains unchanged despite the market’s AI jitters from DeepSeek, and this was “evident by the significant increase in [the] customer’s recent trial production volume forecast, indicating an accelerated timeline for CPO technology to enter mass production.” They also expect sample volume increases over the next few quarters. 

Most importantly, Himax’s management outlined what CPO’s revenue contribution could look like in the future, with CEO Jordan Wu emphasizing that Himax is now “more optimistic than ever” about the outlook for WLO and CPO products, which they believe are “poised to generate significant growth opportunities and become a major revenue and profit contributor in the years ahead.” 

Here’s what Wu said about 2025, 2026 and the revenue opportunity in dollar terms: 

“2025 will be a year for engineering validation with only sample shipments for us. So while the sample shipment will accelerate quarter by quarter, the revenue contribution will be rather limited, if you compare that with our total revenue. The fourth quarter, which presumably will be the peak of this year, the revenue is set to be in millions of dollars based on current forecast. But again, it's still quite small compared to our total size and things can still change. It's still early.sample shipment will accelerate quarter by quarter, the revenue contribution will be rather limited, if you compare that with our total revenue. The fourth quarter, which presumably will be the peak of this year, the revenue is set to be in millions of dollars based on current forecast. But again, it's still quite small compared to our total size and things can still change. It's still early. 

Now, in all likelihood, mass production will commence in 2026, but we don't know how, and certainly we won't comment on exactly when, in 2026. It's probably still more than a year from now. There are still many unknowns, like how many customers, how many projects, or their ramping curve, etc. And therefore, while 2026 is likely to be the first year of mass production, it is still early and it's still difficult to give a revenue indication for the year at this point.in all likelihood, mass production will commence in 2026, but we don't know how, and certainly we won't comment on exactly when, in 2026. It's probably still more than a year from now. There are still many unknowns, like how many customers, how many projects, or their ramping curve, etc. And therefore, while 2026 is likely to be the first year of mass production, it is still early and it's still difficult to give a revenue indication for the year at this point. 

Now, if we look further ahead and ignore the exact timing and ramping curve, etc, for the time being, and just try to kind of paint a picture for, I would call it annualized potential revenue for Himax,… when the CPO business reaches an ‘early stage’ of mass production, right … when the industry is perhaps still testing the water, with maybe only premium models equipped with CPO.  

In making the assessment, we have considered leading AI customer total advanced GPU shipment outlook … and the leading foundry's total CoWoS capacity plan, which is very much public information. And we have assumed a very low percentage of CPO attach rate for both, right. … So with such conservative assumptions, I can say that our annualized CPO revenue could still reach hundreds of millions of dollars when we get there. Again, this is early stage, and this is the best I can do in terms of providing a revenue indication.” I can say that our annualized CPO revenue could still reach hundreds of millions of dollars when we get there. Again, this is early stage, and this is the best I can do in terms of providing a revenue indication.” 

Wu believes that the determining factor is not a question of whether CPO adoption will occur, but rather a question of how fast CPO will penetrate the industry due to the benefits it offers to power consumption, bandwidth, and cost. Wu added that he also believes the ultimate demand for the new tech is likely to be far above what Himax can predict at this stage. 

Putting Hundreds of Millions of Revenue in Perspective 

While hundreds of millions of dollars is but a mere splash in the bucket for a company like Nvidia or TSMC when it comes to AI-driven revenue, that opportunity is remarkably large for a company like Himax. 

Himax reported just $907 million in revenue in 2024, or just over a (4%) YoY decline due to global demand weakness and conservative purchasing trends from customers due to market uncertainty. Gross margin improved more than 2.5 points YoY to 30.5%, its first annual expansion since 2021, aided by strong 20% YoY growth in the automotive segment (50% of revenue in Q4) which Himax says enjoys a higher gross margin than its corporate average. 

Himax’s revenue performance has been challenged since 2021, where the pandemic-driven operating environment led to demand outpacing supply in core display driver end markets, leading to 110% YoY automotive revenue growth, 77% YoY tablet IC growth, and 85% YoY smartphone growth. Himax generated $1.55 billion in revenue in 2021, up more than 74% YoY while gross margin nearly doubled to 48.4% due to very favorable price and product mix. Revenue has declined YoY each year since then. 

For an emerging new opportunity that has yet to generate meaningful revenue, WLO holds the potential to grow into the size of 25% to 50% of Himax’s current annual revenue, or multiple times the size of its non-driver business, which has reported revenue between $140 million to $185 million the past four years.  

The CPO-driven opportunity holds remarkable potential to dramatically increase Himax’s topline and growth and put it on a path to quickly reach fresh records for revenue. However, this opportunity also extends down the line as it may provide a substantial boost to Himax’s earnings power.  

Himax is currently operating at around a 30% gross margin and high-single digit operating margin profile. However, this operating margin is driven entirely by Himax’s driver segments, as its non-driver segment had generated widening operating losses from 2021 through 2023 before slightly rebounding in 2024. 

Operating losses for Himax’s non-driver segment widened from ($17.7 million), or an (11%) margin, in 2022 to ($32.1 million), or a (23%) margin, in 2023. However, losses rebounded slightly to ($24.5 million) in 2024, or a (16%) margin. This weakness amplified the rapidly shrinking operating income from drivers, which faced pricing pressures in 2022 and 2023, and weighed down on the segment’s rebound to $93 million in operating income in 2024. 

Himax explained that low sales volumes in non-drivers “led to insufficient revenue to fully cover expenses” over the past few years, but as volume production ramps, such as for products including WLO and WiseEye, Himax expects to generate positive operating income as non-driver products have “higher gross margins as well as higher growth potential” versus drivers. 

Given some of the uncertainties about timing of CPO’s ramp, volumes at mass production, realized selling prices and other factors, Himax has provided no insight into what the margin profile for WLO could look like. Assuming the CPO business ramps as expected and matures into a 40% gross margin, 15% operating margin business at an $500 million dollar scale, it could generate $200 million in gross profit and $75 million in operating income by itself. This alone could represent a ~$0.40 positive impact to EPS.  

If CPO matures into a much higher-margin business due to Himax’s and FOCI’s positioning in the supply playing a core role to help foundry partners meet high demand, say at 60% gross and 30% operating margins (approx. in line with 2021’s corporate operating margin), it could generate $300 million in gross profit and $150 million in operating income, or up to $1 in positive EPS impact.  

At the moment, there is little analyst coverage and limited visibility into Himax’s growth prospects. Current analyst estimates point to a slight YoY decline in EPS in 2025 to $0.44, with revenue rising 8% to $981 million as Himax works its way out of a more challenging macro backdrop. For 2026, EPS is expected to double to $0.88 with revenue growth of 20% to $1.18 billion, though this could be impacted by the timing and pace of CPO’s ramp.  

Himax does have other growth opportunities outside of CPO, with management anticipating strong growth in its WiseEye business in 2025 and continuing momentum in automotive TDDIs, timing controllers and growth in auto OLED. 

Quick Note on Technicals 

In 2019, Himax appears to have put in a major low.  That being said, investors should still expect large swings and the potential for bouts of volatility as this larger uptrend pattern plays out.  

The current drop that we are in is hitting strong support just under $8. If this level can hold, we should see one more push into the $17 region in the coming months, which would complete a large 1st wave.  If this does happen, we would then need a 2nd wave pullback to follow.  This would likely take you back into the $10 – $7 range, before commencing with the larger uptrend pattern. If any bout of volatility takes us under $6.40, then we should expect a drop into the $4 – $2 region before finding a meaningful low.  

While the larger pattern suggests much higher prices for HIMX, based on the overlapping nature of this pattern, investors should expect periods of heightened volatility along the way. The $6.40 price point is a major line in the sand. If this holds and we see a push into the $17 region, investors should also be aware for the potential of a notable pullback before taking off.  

Major Risk: China Sourcing: 

Himax operates via a fabless business model utilizing third-party foundries and OSAT capabilities, and primarily sells products via direct sales teams in core regions Taiwan, China, South Korea and Japan. However, Himax generates more than three-fourths of its revenue from China, which could place it in the cross-hairs of increased geo-political tensions.  

Himax notes that it relies on TSMC, UMC, GlobalFoundries Singapore, Macronix, PSMC, Nexchip and SK Hynix System IC for wafer fabrication, while its WLO production is done in-house in its two facilities in Taiwan. 

China’s revenue share declined slightly from 76.2% in 2023 to 73.4% in 2024, with a majority of this revenue (85%) in the Drivers business. China accounted for nearly 77% of Himax’s Drivers business in 2024, and 63% of its Non-Drivers business. While Himax may not be exposed directly to US-based tariffs due to its concentration in Asia for supply chain and manufacturing, it warned that it expects to “continue to be subject to economic and political events and other developments that affect our customers in Asia,” and if tariffs dent consumer electronics demand either in Asia or globally, Himax could face revenue headwinds as a result.  

The I/O Fund owns a stock that is rumored to be a future CPO supplier to Nvidia (and is a current Nvidia supplier) while being vertically integrated on sourcing. This company’s CEO pointed out this is an area incremental strength and an advantage as they manufacture most of their parts internally (which is quite rare) and can quickly capitalize on a rapid ramp in demand. The company stated that they also will source when the demand requires it, yet the vertical integration on the manufacturing side with primary production in the US, and some in Europe, help avoid heightened tariff risks from China.

Conclusion 

CPO beckons a massive opportunity ahead for Himax due to its positioning upstream in the supply chain for TSMC. Management outlined that the emerging technology could represent hundreds of millions of dollars in annualized revenue assuming a low attach rate as the tech ramps through 2026 and beyond.  

While CPO promises hundreds of millions of potential revenue and possibly a large positive impact to EPS as it matures, Himax’s driver business remains in the driver’s seat for 2025 and 2026 until CPO growth arises, and the segment is still plagued by margin weakness. The driver business is far from 2021’s peak for revenues and margins, and thus Himax’s EPS, and it may struggle to return to 2021’s levels due to the supply chain circumstances and resulting pricing advantages it benefited from at the time.   

Despite that, Himax appears relatively priced ahead of a potential revenue acceleration and development of a multi-million dollar business line. Himax is currently trading at 18x forward earnings, with EPS estimated to double in 2026, and 1.4x forward sales, with prior peaks around 3x. Gross margins are showing strength due to strong automotive sales and strict cost management, while operating income is beginning to recover from a prolonged contraction from headwinds in the driver segment.  

Small caps can have quick and volatile moves, and Himax is no exception. Shares surged 45% in one day in December when it was first named as a potential AI supplier to TSMC, and then rose nearly 100% in seven days in January before plunging nearly -28% during DeepSeek’s market rout and is down nearly 50% from the January high. 

The I/O Fund owns a different NVDA and possible future CPO supplier, sharing this research with our Pro and Advanced subscribers, while discussing potential setups and trading plans in our weekly webinars with Portfolio Manager Knox Ridley. Take advantage of a limited-time offer for $75 off Pro or $100 off Advanced here.

Please note: The I/O Fund conducts research and draws conclusions for the Fund’s positions. We then share that information with our readers. This is not a guarantee of a stock’s performance. Please consult your personal financial advisor before buying any stock in the companies mentioned in this analysis.

Recommended Reading:

Essentials Report, April 2025

Posted on by io-fund

We recently published a free article that outlines the major risks this market currently faces. We also provided broad market analysis on the two most likely paths this market might take. For those interested, we encourage you to read it herehere.  

Regarding the broad market, in our last report, we stated that  

“The line in the sand is 5400… If the market decides to break below 5400 in direct fashion without a bounce, I think it is reasonable to expect a more direct drop to 5200 – 4800.”The line in the sand is 5400… If the market decides to break below 5400 in direct fashion without a bounce, I think it is reasonable to expect a more direct drop to 5200 – 4800.” 

We found a low, so far, in the 4800 region, as the market is now staging a bounce. With sentiment readings at historic lows, if the market is going to push lower in a bigger way, it will need to reset sentiment and bring investors back into the markets. This further supports a bounce, which we believe has already started. There are two basic paths I’m tracking over the short-term time frame: 

  • Blue – We have started the expected bounce, which should take us to at least 5635 and potentially to the 6050 region. I think we will see some weakness this week that should hold over 4980 SPX. If this level holds, we should see the bounce continue to our targets.

    Once we get to the target region overhead, risk will be elevated. How we correct from there will be crucial, which will determine if we are heading below 4000 SPX, or if we are setting up for a move to new all-time highs.  

  • Red – Any further weakness breaks below 4980 SPX and continues lower. In this instance, I will be targeting the 4700 – 4546 to complete this leg of the drop. This should be the last swing lower that will set the stage for a multi-week bounce back into our overhead target zone.   

As stated in the last report, we reserve regular broad market updates for our Advanced Members. However, during times of volatility, which could lead to a meaningful inflection point for investors, we want all our members to have the proper context so that they can risk manage their positions.  

If you are interested in a day-by-day update on both broad market risk, and how the I/O Fund manages market volatility in real-time, please consider our Advanced Tier. We are offering a 30% discount to all Essentials Members – please contact support@io-fund.com for details.. We are offering a 30% discount to all Essentials Members – please contact support@io-fund.com for detailssupport@io-fund.com for details.

Nvidia (NVDA) 

The bounce in NVDA looks a lot like the S&P 500 overhead – weakness into this week that makes a higher low, then a more direct push into the $116 – $127 region. The drop looks incomplete, and there are two scenarios I’m tracking on what could potentially happen after this bounce completes: 

  • Blue – This bounce will fail to make a new high, and the final drop that follows will complete a very long correction in the $70 region. Once the coming bounce completes, the drop that follows needs to be a messy/3 wave move for confirmation.  In other words, we do not want to see a direct drop with bounces.  
  • Red – This scenario should look identical to the blue scenario – a bounce into the $116 – $127 region. However, once this bounce completes, we’ll see a more direct drop, signaling that we are setting up for a move to the $60 region.  

Taiwan Semiconductor (TSM)

TSM has a similar setup as NVDA and the S&P 500 above, and is setting up for a big bounce, which has either already started or will start after one more slight low.  We expect TSM to take us back to the $187 – $215 region on this bounce. Like with all the equity positions we track, once we get into the target zone, risk will be elevated. 

We will not know if the coming bounce will be a lower high in an even larger correction, or a pause on our way to new highs. Based on the mounting risks, we may be looking to reduce risk in TSM and many other positions on this bounce.  

Bitcoin (BTCUSD) 

Bitcoin is moving in a different path than the equity markets. Bitcoin now has all waves in place to suggest this correction is over. However, I do believe that the pattern best fits with one more slight low to the $74,000 – $69,000 region. If this drop lower happens, it should be on less momentum and less volume than prior drops. This will signal the final 5th wave lower is in place, which should then complete the larger correction.  

If instead, we can see price break over $88,500, then the odds will favor the correction being over. It will be a series of steps, but the larger swing, which we have been anticipating for many months, should take us into the $120,000, at minimum. If this plays out, we will greatly reduce our Bitcoin exposure and continue to log meaningful gains.

Please note: The I/O Fund conducts research and draws conclusions for the Fund’s positions. We then share that information with our readers. This is not a guarantee of a stock’s performance. Please consult your personal financial advisor before buying any stock in the companies mentioned in this analysis.

Recommended Reading:

The Fed Can’t Save This One: Why Bonds May Break the Stock Market in 2025

Posted on by io-fund

Last February the S&P 500 was nearing all-time highs and exuberance could be felt throughout the market with crypto leading the way. Going against popular opinion, we stuck our neck out in the analysis: “AI Stocks Signal a Correction Before a Buying Opportunity Emerges” to state:

“While the market continues to push higher, it has been doing so without the support of key stocks and important sectors. As we’ve seen in times past, while this divergence can go on for a while, unless it invalidates with all sectors and stocks breaking to new highs, it tends to act as a warning. With money managers and retail investors all-in on stocks, it appears that the stage is being set for a potential rug pull this year.”

We went on to say that we have set up “aggressive buy targets” in key AI names, and that this period of volatility could set up a great buying opportunity. February wasn’t the first that we discussed this strategy, rather on January 1st – before DeepSeek or tariffs, we stated Nvidia’s next price target was between $102 – $83, causing us to cut more than half pf our position at $140 and $126 in February.

This past month, we have been following our buy plan, using some of the 50% cash we raised in December of 2024, as we expect a sizable bounce. From what we know today that bounce should be used to de-risk, although this ultimately needs to be assessed when we reach key levels.

With all major indexes in a technical bear market, and sentiment readings at historically low levels, most readers want to know if this is a generational buying opportunity, or are we setting up for lower levels in the coming weeks to months? Sentiment suggests that, even in the worst-case scenario, we should see a sizable bounce. However, with the shifting geo-political landscape, and especially the reaction in the bond market, we have shifted our stance to using the coming bounce to de-risk as there is the potential for a much larger bear market than what is currently priced in.

Why Bonds Matter 

The trade war is taking center stage, and for good reason. Liberation day shocked the markets in the scope and severity of tariffs imposed on foreign nations, leading to one of the most extreme 3-day drops in market history. The tariffs and what this will mean for earnings and domestic growth is causing a level of uncertainty rarely seen in public markets.

While the masses are focused on the potential resolution and what will happen if we do not see one, there is a more immediate problem – no one is buying government bonds despite growth slowing down in the economy. This is not normal behavior and will become a problem if not resolved. Yields are too high for new bonds being issued, whereas the US government needs lower yields to finance the breathtaking amount of debt coming due this year.

In our October report, we quickly pointed out the unusual reaction bonds had to the FED’s surprise 50 bps cut. Bonds began dropping with rates, which are continuing to play out into today. Historically, bonds tend to go up (yields down) when the FED begins a rate cutting cycle.

The rationale is that the FED is cutting because they see growth and inflation slowing. If demand slows in the economy, prices drop to meet that demand, both in goods and services, as well as stocks. This is the type of environment investors would want to own a fixed yield coupled with the perceived safety found in Treasuries.

Join thousands of investors who trust I/O Fund’s expert stock analysis on AI, semiconductors, cryptocurrency, and adtech — sign up for free! Click here!Join thousands of investors who trust I/O Fund’s expert stock analysis on AI, semiconductors, cryptocurrency, and adtech — sign up for free! Click here!

This logic is also why bonds went into a bear market in 2022 with stocks. In a rising inflation environment, where prices are going up, owning a fixed yield is not ideal, as the purchasing power of that fixed yield will diminish each year.

Aside from 2022, the inverse correlation between stocks and bonds has been a market axiom dating as far back as 1998. When growth is expected to slow, stocks tend to peak, and bonds go higher.

2008 

Treasury bonds rose as stocks fell in 2008, showing a 25-year safe haven trend.

Treasuries bonds moved higher as stocks moved lower in 2008. This has been a 25-year pattern, as investors seek a safe, fixed yield in an environment where prices go lower.  

2010

Treasury bonds rose as stocks fell in 2010, reflecting a long-term safety trend.

Treasuries bonds moved higher as stocks moved lower in 2010. This has been a 25-year pattern, as investors seek a safe, fixed yield in an environment where prices go lower. 

2015

Treasury bonds rose as stocks declined in 2015, continuing a 25-year safety trend.

Treasuries bonds moved higher as stocks moved lower in 2015. This has been a 25-year pattern, as investors seek a safe fixed yield in an environment where prices go lower. 

2018

Treasury bonds rose while stocks fell in 2018, continuing a 25-year flight to fixed-yield safety.

Treasuries bonds moved higher as stocks moved lower in 2018. This has been a 25-year pattern, as investors seek a safe, fixed yield in an environment where prices go lower. 

2020

In 2020, Treasuries rose while stocks declined, reflecting a 25-year trend of investors seeking safe yields.

Treasuries bonds moved higher as stocks moved lower in 2020. This has been a 25-year pattern, as investors seek a safe, fixed yield in an environment where prices go lower. 

Regarding today’s environment, we are seeing all major indexes in bear market territory, while GDP projections are trending deeper into negative territory. The Atlanta GDP Now forecast went from +2% forecasted for Q1 of 2025 to -3% in just under 2 months.

Evolution of Atlanta Fed’s GDPNow real GDP estimate for Q1 2025, compared to top and bottom Blue Chip forecasts.

Atlanta GDP NowAtlanta GDP Now 

Furthermore, Goldman Sachs raised their odds of a recession in 2025 to 45%, while the odds on most major betting platforms have a recession at 60%. These odds have backed off somewhat due to the erratic news surrounding the Trade War, yet they remain uncomfortably elevated by historic standards.

This is being confirmed with the Oil markets breaking down from a 3-year range, which has moved sharply lower and appears to be confirming a new downtrend. As the global demand for oil is fading, this is signaling the likelihood of a notable slowdown in global growth, not just in the U.S.

Oil price breaks down from a 3-year consolidation, signaling increased risk of a global recession in 2025.

Oil is breaking down from a 3 year consolidation, signaling the rising risk of a global recession. 

By all accounts, we are in a slowing growth environment, with stocks and oil moving vertically lower into a bear market. If there is any environment in which you’d expect bonds to catch a bid, this would be it. However, since the February high in stocks, TLT is -1%, meaning that the historically safe have of Treasury Bonds are not in demand. This may change, but based on history, we tend to see Treasury Bonds up a meaningful amount this late into a growth slowdown cycle.

For those paying attention, something is different in this cycle. A 25-year correlation between stocks and bonds appears to be breaking. While many are getting caught up in trying to explain why this is happening, the why is simply not important right now. What is important is understanding what the ramifications of this broken correlation will mean in 2025, and if this persists, we could see a more immediate problem in equities than the trade war. 

Stocks and bonds’ 25-year correlation breaks in 2025, as recession risk rises and demand for treasuries falls.

The 25-year correlation between stocks and bonds is not holding up in 2025. Considering the rising recession risk, and a technical bear market is stocks, the safe, fixed yield offered by treasuries is not in demand. 

The Government and FED Needs Lower Bond Prices

As of December 2024, The United States debt is currently 124% of it’s GDP.  This was before GDP was forecasted to shrink in Q1, while Debt is forecasted to continue to rise, even with the efforts of DOGE. This will certainly bring the U.S. into the dreaded 130% Debt/GDP region. Since 1981, 98% of all countries that have reached the 130% Debt/GDP ratio have defaulted on their bonds.

Japan avoids default with 130% debt to GDP, but a default remains likely, as predicted in the 2016 letter.

Bloomberg

The problem the US now faces is a runaway bond market, much like we saw in England in 2022. In other words, as new debt is issued with higher yields to pay, more debt will have to be issued to service this debt – i.e., pay the yields. As more debt is used to service the existing and new debt, this puts more money into the economy, which is ultimately inflationary.

As we already know, bonds don’t like inflation, so they go lower, pushing yields higher. This causes a spiral effect. The bond market demands a higher risk premium (higher yield) to hold bonds to maturity, which causes the government to issue more bonds to service this rising risk premium.

I believe the market is signaling to the investors that this process has begun. To put this into perspective, the budget deficit for the fiscal year 2024 came in around $1.9 Trillion, or 6.7% of GDP. There is no other year in US history where the budget deficient was this large outside of a major war, like WW I & WW II, or dealing with a major recession, like 2008. It is unheard of to have fiscal spending this high, in an expanding economy, with historically low unemployment.  

From this excessive level, what will deficit spending be if we enter a recession? The biggest misstep the FED and government can make at this juncture would be pivoting while bonds are going down. Until we see Treasury bonds (TLT) breakout and move higher, the FED and government run the risk of losing control of the bond market in a disorderly manner, which would also further tank equities.

This is what’s concerning right now. While we have been trained to hang on until the FED can swoop in and save us, we are being held hostage by the bond market, which is only reacting to decades of reckless spending finally coming to a head.  

If a 20% drop in equities, vertical drop in oil prices and growing certainty of a global recession is not enough to get bonds off the floor, what will it take? The bureaucrats that created this mess run the risk of losing control of the bond market if they act too soon. They have no choice but to sacrifice the stock market until bonds catch a meaningful bid.

You might be wondering, why does this matter? Why can’t the FED and Government allow Treasury Bonds (TLT) to go lower? Simply put, the U.S. has to refinance $9.2 Trillion in US debt in 2025 with an estimated $28 Trillion in debt needing to be refinanced over the next 4 years. The cost to refinance this debt will depend on the yield the market is demanding to hold a US government bond until maturity. The lower ETFs like TLT go, the more of a problem the US will have to refinance this debt, further exacerbating the runaway bond market scenario.

For reference, as of February 2025, it currently costs $478 Billion annually to maintain the US debt. This is roughly 16% of the total Federal Spending. Also, for the first time in history, it now costs us more to service the debt than we spend on defense annually.

[mid-signup-palaceholder]

The FED does not control the bond market, though they would like you to think they do. The closer you get to the front end of the yield curve – i.e., 1-month yields of T-Bills to 2-year yields on T-Bills, the more of an influence the FED has on these bonds. However, when you start getting into the 10-year yields and farther out on the yield curve, the more of an influence growth and inflation outcomes have on yields. The bond market is saying, we don’t believe you’ll pay this back, and the consequence of this spending will only lead to inflation. Therefore, we want a higher risk premium (yield) to hold these bonds until maturity. The only problem with this is that the U.S. can’t afford higher yields and must refinance 24.5% of its national debt this year alone.

Sentiment

Currently, the stock market is reacting to the uncertainty in the global economy. What could cause this uncertainty to move lower is how bonds are not moving higher. This could lead to further selling; however, sentiment readings are suggesting that a sizable bounce should happen before we are going to move meaningfully lower.

Investing is a zero-sum game. In other words, for you to win, others must lose. For this reason, measuring what the masses are doing can be quite beneficial as a contrarian indicator. For example, in our February report, we discussed in detail how both fund managers and retail investors were all in on stocks. Almost every measurement used to gauge positioning was at or close to new historic extremes. We tend to see these extremes at meaningful turning points, which is exactly what happened.

Now, with all major US markets confirming a bear market in just under 2 months, we are seeing extreme sentiment readings on the bearish side of the equation. For example, the AAII Investor Sentiment Survey asks retail investors where they think the market will be in the coming months. It has been around since the 1970s and has a remarkable history of signaling turning points in the market through sentiment extremes.  

Last week marked the 6th consecutive week in a row where the bearish sentiment was in 97th percentile or higher of all weekly surveys going back to the 1970s. There was no period on record that matched this level of extreme consecutive bearishness – not the COVID lows, the 2009 lows or the 1990 lows.  

AAII investor sentiment hits most bearish 6-week reading in history, with 97th percentile bearish readings for 6 consecutive weeks, often signaling a contrarian indicator.

The AAII investor sentiment survey is showing the most bearish 6 week reading in its history. This is the 1st time since the 1970s that we have seen bearish reading in the 97th percentile or higher for 6 consecutive weeks. This data is typically used as a contrarian indicator. 

With investors feeling this bad about the markets, history suggests that we should see a sizable bounce, at minimum, soon.  Just like everyone piling into stocks at the top, we tend to see capitulation to match extreme bearish sentiment. With the Volatility Index (VIX) hitting 52, which is higher than any point in 2022, as well as the S&P 500 losing 15% in three days, it appears that we are getting capitulation or are close to capitulation.

VIX spikes to highest levels since COVID crash, with a 15% drop in equities, signaling potential capitulation.

The VIX is has spiked of the highest readings since the COVID crash, coupled with a 15% drop in equities in only 3 days. Could this be capitulation? 

What investors need to see is the VIX settle back below 30, and the recent lows made this week hold for confirmation. If both conditions are not met, we expect another leg lower in this portion of the bear market before a bounce.

Broad Market Analysis 

Anyone who has been following along based on our analysis should not be losing sleep over the current bout of volatility. Based on the constant warnings we provided throughout 2024, we were in more than 50% cash in January of 2024, and even moved into a 100% hedge position in February of this year.

While we see the potential for another leg lower in this bear market, we should see a sizable bounce first.

We have consistently stated that the first major shot across the bow for the markets is if the S&P 500 breaks the 5400 region. We found support around the 4835 region. If this level breaks, the next level lower is in the 4600 region. With this information accounted for, below are the two scenarios I see as most likely given the current price action.

Red – We are completing the 1st leg down in a much larger bear market. The next move will be a corrective rally that makes a lower high. The targets for this bounce are between 5600 – 6050. This will be followed by the final move lower in the bear market toward 4200 – 3500 SPX.

Green – This drop holds over 4260, which will complete a 4th wave decline in a very large ending diagonal pattern that started at the 2022 lows. The next move higher will take us to the 6300+ region, which will be the final 5th wave in the bull market that started in 2022.

S&P 500 bear market scenarios: all-time highs by fall 2025 or a lower high/bounce before another leg down.

Two Scenario that this bear market in the S&P 500 can play out. Either we put in a big low, and can see all-time highs by fall of 2025, or we will see a lower high/bounce into June that will lead to another leg lower in this bear market. 

Further supporting the red scenario, the S&P 500’s Relative Strength Index (RSI) has broken below standard bull market support. 

The RSI tends to find support around the 40 region in bull markets and breaks above the 60 – 70 region. In bear markets, the RSI will break below bull market support and fail under the 60 region in bear market rallies. The fact that we broke the defined bull market support for the S&P 500 is concerning.

The first market to break its RSI bull market support was the Russell 2000 (small caps) in mid-February.

The Russell 2000 small-cap index is the first to signal the upcoming bear market in 2025.

The Russell 2000 small cap index was the 1st index to warn of the coming bear market in 2025. 

This was followed by the NASDAQ-100, which broke its bull market trendline along with its RSI bull market support in early March. 

The NASDAQ-100 Big Tech Index was the second, and most important, market to signal the upcoming bear market in 2025.

The NASDAQ-100 Big Tech Index was the 2nd, and most important, market to warn of the coming bear market in 2025. 

The S&P 500 just broke its RSI bull market support last week, as it found support around the 4835 region, which typically signals a meaningful shift in momentum. If the next large bounce fails under the 60 – 70 region, it could be an early warning that the red scenario is underway.

Conclusion:

In conclusion, we have been warning investors, even in the bull market of 2024 that the uptrend is unhealthy. The many divergences and long-term Elliott Wave patterns were not as healthy as the trend suggested. We further warned our readers to expect more volatility in the February report – at the time, a lone voice in the sea of exuberance. The market has now retraced the entirety of 2024’s gains in less than 2 months.

While we have been buying beaten down AI stocks around these lows, we may not hold them too long into 2025 if the coming bounce starts to fail, further signaling the red scenario is in play.  

Investors have been trained to buy-the-dip. Every bear market since 2009 has recovered in short order. Even the COVID decline saw one of the quickest bear markets, and quickest recoveries in history. The difference between these periods and what is happening today is twofold: 1) we never saw a real recession, where growth slows for several quarters and stays at depressed levels; 2) The FED was allowed to put a floor under the stock market with excessive liquidity because inflation and bonds allowed this to happen.

Investors need to be prepared for a changing dynamic, which is being signaled by the bond market. As always, we are buying stocks at depressed prices in hopes of the best-case scenario, but we continue to prepare for the potential of a worst-case scenario.  

Regardless, AI will continue to innovate, as we are in the 1st inning of this multi-decade tech trend, and stocks will eventually find a floor, setting up generational buying opportunities for those prepared.

If you went into this sell-off fully invested without any risk management plan, or if you are sitting on outsized losses and not sure what to do, we encourage you to attend our upcoming weekly webinar for premium members. Next Thursday, April 17th, at 4:30 ET. In this upcoming webinar, we will discuss our game plan regarding the remainder of 2025. We will list buy targets for great AI names as well as go over how we plan to raise cash and further hedge our portfolio if this bear market continues into 2026.

The I/O Fund is a leading tech portfolio with annualized return of 27.6% — which would rank us as #2 in the United States if we were a hedge fund. Learn more here.Learn more here.

Please note: The I/O Fund conducts research and draws conclusions for the Fund’s positions. We then share that information with our readers. This is not a guarantee of a stock’s performance. Please consult your personal financial advisor before buying any stock in the companies mentioned in this analysis.

Recommended Reading:

Meta Platforms: Quietly the Strongest Mag 7 Stock with Meta AI App Launching Soon 

Posted on by io-fund

Meta is the strongest Mag 7 stock YTD with its peers down as much as 40%, Meta’s (13%) decline since the first of the year is strong on a relative basis and provides some important clues. 

It may be easy to lump the Mag 7 together, yet this cohort is different in key aspects that matter. For example, we’ve been cautioning on Tesla’s margins for years, Nvidia continues has an AI lead that appears insurmountable, while the others have open pocketbooks for AI chips without much visibility to offer in terms of how that capex will be paid back and when.  

It’s this last piece that Meta has a better handle on.  The company is not offering data center infrastructure for developers and R&D departments to create AI applications, who in turn, must find a path to monetization for a virtuous cycle of demand. With what we know today from earnings disclosures, Azure is leading with a $13 billion annual run rate in revenue among the cloud providers (Amazon, Google, Microsoft). Yet, consider that Meta has a new ad product driven by AI that is already on a $20 billion run rate, and this does not include incremental revenue AI is driving across their other ad products, evidenced by an inflection in Meta’s key metric on ad prices. 

There are other areas that Meta is quietly leading. The company is releasing a standalone AI app in Q2, powered by Llama 4, with projections for 1 billion monthly active users by year’s end. That’s up from 700 million users today on the embedded AI features it sees in its popular “family of apps” such as Facebook, Whatsapp and Instagram. To compare, Chat-GPT's standalone app has 500 million. While it remains to be seen if Meta AI will reach the company’s lofty goals, there’s certainly a viable path with 3.35 billion currently using its popular apps today. 

Lastly, whereas other big tech companies ramped capex 1-2 years ago, Meta is only now moving into large capex spending from $30B in 2023, to $40B last year to $62 billion this year, up 59.4% YoY. This past quarter, capex was up an astonishing 88% YoY. Meta is in the enviable position to increase this capex once they’ve found product market fit within their own platform rather than pre-emptively building for future AI development. I suspect this new trajectory in capex spending means Meta is ready to scale its AI tools on the ad platform side.  

Meta’s AI-Powered Ad Tools 

Meta is unique in that its customer base has an unusually high incentive to use AI tools. You could argue developers for cloud IaaS also have a high incentive, yet they must in turn monetize to drive scale for their AI apps. Meanwhile, advertisers see an immediate ROI using AI tools and are then incentivized to spend more.  

Evidence of this is visible in the stubbornly high ad revenue growth last quarter of 20.6%, which has sustained even as Meta recently lapped tough comp as the average ad price grew 14% YoY.  

Advantage+ has $20B Run Rate:

Advantage+ is an automation tool that uses machine learning to match customers to ads. Similar to Google’s Performance Max, the tool uses AI and machine learning to help with optimize bids/budgets, create and find audiences, and to produce ad creatives. Up to this point, Advantage+ has been used for ecommerce and manually turned on, yet Meta is now going to have the tool turned on automatically and roll it out to other verticals soon.  

“In this new setup, all campaigns optimizing for sales, app or lead objectives will have Advantage+ turned on from the beginning. This will allow more advertisers to take advantage of the performance Advantage+ offers, while still having the ability to further customize aspects of their campaigns when they need to. We plan to expand to more advertisers in the coming months before fully rolling it out later in the year.” 

According to the recent earnings report, over 1 million advertisers used Meta's gen AI tools to create 15 million ads in December 2024.  

  • Advantage+ shopping campaigns surpassed a $20 billion annual run rate, up 70% YoY.  
  • Advantage+ users are seeing a return on ad spend (ROAS) improvement of up to 22%. 
  • Advantage+ Creative is seeing over 4 million advertisers using at least 1 generative AI ad creative tool, up 1 million just six months ago.  

Andromeda has 10,000X increase in Model Capacity: 

In H2 2024, Meta revealed a new machine learning system built on Nvidia Hopper chips called Andromeda. Per the earnings call: “This more efficient system enabled a 10,000 times increase in the complexity of models we use for ads retrieval, which is the part of the ranking process where we narrow down a pool of tens of millions of ads to the few thousand we consider showing someone.” 

Consider that with AI, an advertiser can make 5,000 creatives rather than 5 creatives, changing the game in personalization. Andromeda is aimed at the new frontier of AI advertising, with an emphasis on retrieval as the first step. This first stage in the ad serving process is to retrieve the ads that are most personalized to a user, resulting in thousands of ads before the final ads are selected.  

Due to AI, these ads can now be chosen based on learned associations rather than categories or keywords. By offering a 10,0000X increase in model capacity, Andromeda processes large amounts of data to accurately predict which ads will resonate with which users. This leads to more relevance, which in turn, leads to higher conversions. The deep neural network is considered "hierarchal structured” to where it can handle a much larger volume of ads at a much higher accuracy, improving on the previous structure known as “two-tower.” 

Meta AI to See 1 Billion Users in 2025 

In a recent conference, Chief Product Officer Chris Cox, described Meta’s AI features as Search 2.0. Instead of visiting a search engine, users engage with Meta AI in-app and on the newsfeed. Meta’s standalone app is expected to be announced in late April at the developers conference.   

During the conference, Zuckerberg said, “Meta AI differentiates itself in this category by not just offering state-of-the-art AI models, but also unlimited access to those models for free, integrated easily into our different products and apps. So Meta AI is on track to being the most used AI assistant in the world by the end of this year. In fact, it’s probably already there. We’re almost at 500 million monthly actives, and we haven’t even launched in some of the bigger countries yet.” While this version was on Llama 3.2, the new rollout of its standalone Meta AI app will be using Llama 4.0 (detailed below).  

Since then, Meta AI has reported 700 million users on a monthly basis while Chat-GPT is at 500 million weekly active users. This is not comparable, however, since Meta AI is not a standalone app yet. The release will be closely watched as Meta will likely push hard on its large social media base to download the app. 

Chatbot to Aid in Personalization of Content and Ads: 

Meta’s chatbot will also help to aid in further personalization of ads. The data from its chatbot will be stored in memory to help target users. 

“We'll be able to remember certain details that people share in one-on-one chats, for example, and use those details to personalize its responses and then really increasing its ability to deliver great content recommendations and enhance really what makes Facebook and Instagram, so valuable for people today.” 

CFO Susan Li described the Meta AI roadmap in 2025 during the Q4 2024 conference call, stating something similar as to the strengths of Meta’s app over other AI chatbots in terms of personalization and a large context window: 

“And as we look forward to 2025 in our Meta AI road map, we are really focused on doing more to make it feel more personalized. So, I would say some of the most exciting features we're working on include improving sort of the memory dimension of the Meta AI experience. We'll be able to remember certain details that people share in one-on-one chats, for example, and use those details to personalize its responses and then really increase its ability to deliver great content recommendations and enhance what makes Facebook and Instagram so valuable for people today.” 

Notably, it’s unlikely Meta AI is monetized this year, although in future years there are plans for a premium tier similar to Chat-GPT. 

Llama 4 Models Released Last week 

Llama 4 Scout and Llama 4 Maverick were released last week, while Llama Behemoth at 2 trillion parameters is expected to be released soon. These multi-modal LLMs are trained on text, images and video with a mixture of experts (MoE) architecture. MoE distributes a computational load across “multiple experts” (or neural networks) and trains across thousands of GPUs using what is called model and pipeline parallelism. This enables more compute-efficient pretraining, yet the parameters still need to be loaded in RAM, so the memory requirements remain high. 

For Llama 4 Maverick, 400B parameters are stored in memory while only 17B parameters are activated during model deployment, which greatly improves inference efficiency by lowering costs and latency.  

This is similar to the breakthrough we saw with DeepSeek R1, which was a MoE model that was trained on 671 billion parameters stored in memory, yet when the model is served, only 37 billion parameters are active. The overall result is computational efficiency as only the most relevant parameters are activated for a specific task. 

  • Llama Maverick can be run on a single H100 DGX system (8 GPUs) or can be run distributed for inference purposes. The inference cost for Llama 4 Maverick is $0.19 to $0.49 per 1 million tokens compared to Chat-GPT4o at $4.38 per million tokens.  
  • Groq has verified the upper end of this inference cost estimate with Maverick at $0.53. The model leverages 128 experts. According to Meta, Maverick exceeds “comparable models like GPT-4o and Gemini 2.0 on coding, reasoning, multilingual, long-context, and image benchmarks, and it’s competitive with the much larger DeepSeek v3.1 on coding and reasoning.” 
  • Llama 4 Scout is cheap at $0.13 per million tokens according to Groq and can be deployed on a single H100 GPU, leveraging 16 experts. Scout can remember long threads and documents of up to 10 million tokens. This is the largest context window across LLMs available today. 
  • Llama 4 Behemoth is a teacher model that is still being trained with 288B active parameters and 2 trillion total parameters, leveraging 16 experts. Llama 4 Behemoth helped train Maverick and Scout and ranks high on internalinternal benchmarks such as MATH-500, MMLU Pro, GPQA Diamond, Multilingual MMU and image reasoning MMMU. According to MetaAccording to Meta, Chat-GPT 4.5 ranks lower on these benchmarks although this will not be officially verifiedwill not be officially verified until Behemoth is released. For reference, an engineer estimates Chat-GPT 4.5 has 5-7 trillion parameters and 600B active parameters. 

Recently, Chief Product Officer Chris Cox explained the goals for Llama 4 – it's a long quote but also important to hear what Meta is setting out to achieve with its newest models and also as an open-source leader in LLMs: 

“We've finished pretraining the smaller model. The main thing — first, just from an intelligence perspective, we're trying to pack basically, the intelligence of the large Llama 3 series down into really small models, which can then be used with low latency for low-cost on devices on a single host. So basically getting a lot of the intelligence down into the smaller form factor, that's one of the most important things we can do. 

The second is just the basis of what's expected in a new model today, which is reasoning. Agentic use cases, just meaning tool use, ability to use a browser, ability to use other tools. 

And then the third piece is an omni model. So basically, interacting with image and voice natively. So rather than translating voice into text and then text into LLMs getting text out, turning that back into speech, having speech be native. This is a big deal. I believe it's a huge deal for the interface, the product. The idea that you can talk to the internet and just ask it anything. I think we are still wrapping our heads around how powerful that is.” 

Meta Increases Capex 88% YoY in Q4, will increase 59.4% this Fiscal Year 

Q4 capex was $14.8 billion, driven by servers, data centers and network infrastructure investments. Meta raised its 2025 capex guidance to $60 billion to $65 billion, with a midpoint of $62.5 billion, up 59.4% YoY, driven by investments in AI infrastructure and its core business.  

Meta is also extending the lives of its servers and networking equipment. They will use non-AI and AI servers for longer periods of time before replacing them, which will be 5.5 years, resulting in annual capex savings and depreciation expenses, which is included in its guidance. 

In December 2024, Meta announced plans for its 23rd and largest data center in the United States, a 2GW+ AI data center “so large it would cover a significant part of Manhattan” in Richland Parish, Louisiana. The data center plans to outdo Elon Musk’s xAI’s Colossus supercluster of 1 million NVDA GPUs, with plans for 1.3 million GPUs used to train its Llama AI models.   

The 4 million square foot campus will be set on 2,250 acres on the former Franklin Farm site. They will build up to nine buildings and plan to bring 1 GW online by the end of 2025. The site will total over 2GW at full buildout as construction is scheduled to continue until 2030.    

Meta has a co-location deal with Entergy to develop a 1.5GW natural gas power plant located adjacent to its proposed $10 billion data center. Entergy is investing $6 billion in electric infrastructure, including a 10,000-acre solar farm, three combined-cycle combustion turbines (CCBT) split into two sites totaling 2.26GW and over 100 miles of new transmission lines. Co-location refers to deals where data centers are located next to power plants generating electricity to feed directly to the customer and usually bypassing the electric grid, which would be a back-of-the-meter arrangement.  

Meta is seeking approval from Louisiana OSC to begin construction within 10 months. Meta will be matching its electricity use with 100% clean and renewable energy and has pledged to build or acquire 1.5GW of solar power elsewhere in order to offset emissions from the gas plant. Meta will also contribute to a carbon capture and storage project at the Entergy Lake Charles 994MW gas plant. The gas plant is estimated to open between 2028 to 2029.  

Incidentally, a report by The Information claimed Meta is in talks to build a $200 billion data center. However, Meta has not confirmed the rumor. 

Meta Plans to Increase ASIC Usage 

Meta training and inference accelerator (MTIA) is Meta’s custom silicon and is primarily used for inference today for ads and organic content. The goal is to use MTIA for training next year. According to Tom’s Hardware, the “plan is to gradually increase usage if the chip meets performance and power targets."  

There have been delays in the past on Meta’s custom silicon program with MTIA originally launching in 2020, yet Meta had to halt the program to buy Nvidia’s GPUs over the past few years with MTIA v1 launching in 2023. This year, MTIA v2 was launched using RISC-V cores that lowers the overall cost as there are no licensing fees. You can read more about RISC-V, the open-source competitor to Arm in previous coverage here. 

Regarding DeepSeek, Meta confirmed "that doesn’t mean you need less compute,” going on to explain that a new trend is to apply more compute at inference in order generate a higher level of intelligence and higher quality. “I think that's generally an advantage that we're now going to be able to provide a higher quality of service than others, who don't necessarily have the business model to support it on a sustainable basis. 

Financials:

Average Price Per Ad is Inflecting 

Q4 revenue rose 20.63% YoY to $48.39 billion, beating consensus analyst estimates of $46.99 billion by $1.39 billion or 2.96%. Full year 2024 revenue rose 22% YoY to $164.5 billion. 

Management guided Q1 2025 revenue of $39.5 billion to $41.8 billion, with a midpoint of $40.65 billion representing 11.5% YoY growth at the midpoint. This was in line on a constant currency basis, yet came in under $41.46 billion consensus analyst estimates due to 3% FX headwinds representing 13.72% YoY growth.

Revenue growth in the quarter was driven by 21% YoY revenue growth in its Family of Apps to $47.3 billion, of which ad revenue rose 21% YoY to $46.8 billion, comprising 97.57% of total revenue. Other revenue for its Family of Apps rose 55% to $519 million, driven primarily by WhatsApp Business Platform. 

The number of ad impressions served across its services rose 6% YoY, and the average price per ad rose 14% YoY. Ad revenue growth by geographies was led by the Rest of the World at 27%, Asia-Pacific at 23%, Europe at 22% and North America at 16%.    

The average price per ad was up 14% YoY – which is the most important metric to watch as pricing should increase with the AI advancements described above. Growth in average price per ad is already inflecting when you consider Q3 was 11% growth and Q4 of last year was only up 2%. 

Regarding this inflection, the CFO hinted CPMs (cost per 1,000 impressions) may continue to grow: “Overall, we are seeing healthy cost per action trends for advertisers for whatever is the action that they are optimizing for. And we believe we'll continue to get better at driving conversions for advertisers. And when we do, that will have the effect of continuing to lift CPMs over time, because we're delivering more conversions per impression served, resulting in higher value impressions.” 

The Family of Apps average revenue per person (ARPP) rose 15.6% YoY to $14.25, which was a slight YoY improvement from Q4 ARPP growth of 15.4% — so again, consider this growth is being reported on tough comps. 

EPS Grows YoY and QoQ 

Q4 GAAP EPS rose 50.47% YoY to $8.02, beating consensus analyst estimates for $6.76 by $1.26 or 18.67%. This was a sequential improvement from Q3 GAAP EPS, which rose 37.34% YoY to $6.03, beating consensus analyst estimates of $5.29 by 13.99%.  

It’s no secret that AI agents can write excellent code, which can help drive efficiencies at companies by cutting back on the R&D department overhead. According to the earnings call, Meta expects AI to be as good as mid-level engineer, which will positively impact margins and the bottom line: 

“I also expect that 2025 will be the year when it becomes possible to build an AI engineering agent that has coding and problem-solving abilities of around a good mid-level engineer. And this is going to be a profound milestone and potentially one of the most important innovations in history, as well as over time, potentially a very large market. Whichever company builds this first I think is going to have a meaningful advantage in deploying it to advance their AI research and shape the field. So that's another reason why I think that this year is going to set the course for the future.” 

Margins: 

Meta’s margins and income are impressive at levels not seen since the company was the defacto Wall Street darling in 2017.  

  • Q4 gross margin was 81.7%, falling slightly from Q3 gross margin of 81.8%. 
  • Operating margin steadily climbed to one of its highest levels (ever) at 48.3%. Looking back, it was Q4 of 2017 when Meta last reported a higher operating margin. This is up from an OM of 41% last year. 
  • Net margin of 43.1% is similar – one of Meta’s highest on record, up from a margin of 35% last year. 

Cash and Debt Close 2024 at Their Highest Levels for the Year 

Q4 operating cash flow reached its highest level for 2024 at $27.99 billion with a margin of 58%. This is up from $19.4 billion for a margin of 48.4%.  

 Free cash flow fell sequentially to $13.15 billion compared to last year at $11.5 billion. The roughly $14B difference in OCF and FCF is from high capex spend. 

Q4 cash and cash equivalents rose 9.75% QoQ for $77.81 billion in cash. Debt was $28.82 billion.  

Conclusion: 

Clearly, AI is extremely nascent in terms of what it can do with non-stop development and progress coming from tech’s largest players as well as startups. However, also consider that very few companies have a 3B+ global user base to convert to an AI app.  

Meta is positioning itself to become a dominant player in the AI landscape led by its Llama 4 models and standalone Meta AI assistant app. AI has enhanced its core advertising business with ad pricing reaching an important inflection point. It’s also interesting to consider its AI tool Advantage+ has surpassed Azure in annual run rate at $20B compared to $13B. 

The market is rife with noise, it is hard to know what to pay attention to right now. You can expect us to build a strong pipeline of opportunities to seize when the timing is right.

Please note: The I/O Fund conducts research and draws conclusions for the Fund’s positions. We then share that information with our readers. This is not a guarantee of a stock’s performance. Please consult your personal financial advisor before buying any stock in the companies mentioned in this analysis.

Recommended Reading: