When looking at Nvidia’s forward estimates, what stands out is that revenue growth will peak this quarter at 239%. This can be a tricky place for a tech investor when what’s ahead is slowing growth.

Nvidia could raise and beat, as it’s had a penchant for doing lately, but the slowing growth will eventually catch up to the stock and analysts are pegging H2 for this to happen. To contrast, Nvidia will top over the next two quarters according to revenue growth rates whereas AMD is bottoming.

It’s unclear how much of Nvidia’s expected $100 billion for the data center in FY2025 is priced in. This has been discussed since at least the last quarter’s earnings report, yet the valuation is still very reasonable. We look at this and more below to prepare you for the most anticipated earnings report of the quarter. 

Revenue and Earnings:

Nvidia is expected to report revenue growth of 239.4% for revenue of $20.54 billion for fiscal Q4 ending in January. These estimates have been steadily rising since the H100-related historic quarter last May. Last spring, the January quarter was expected to report 40% growth, and by November, the January quarter estimates were at 195%. I want to paint a picture for why Nvidia’s price action has been so strong – these revised estimates create ample room in the valuation.

For next quarter, estimates are for 204.94% growth for revenue of $21.93 billion.

• Fiscal year 2024 ending in January is expected to report revenue of $59.3 billion for growth of 119.7%.
• For fiscal year 2025, the company is expected to report revenue of $94.1 billion for growth of 58.9% with the growth overweight in the first half of calendar year 2024.

Even if we see a beat and raise, the slowing growth in the second half will be hard to overcome due to high comps. As mentioned in the introduction, Nvidia will begin to lap some stellar quarters come the October CY2024 quarter as the growth in October of CY2023 was 205.5% YoY.

Note: See below for bullish scenarios from analysts where these estimates may be too low.

Earnings growth is similar to revenue growth where the current quarter and also next quarter are expected to be stellar.

• Q4 FY2024 January quarter is expected to report growth of 426.4% for EPS of $4.63
• Q1 FY2025 April quarter is expected to report growth of 354.8% for EPS of $4.96
• From there, the July quarter is also strong at 95.4% growth yet tapers off as the company laps the high comps in the October quarter at 41.4%.

EPS growth of 40%+ is nothing to scoff at, yet the exuberance that pushed Nvidia to achieve a market cap of $1.8 trillion to where it is now the world’s second most valuable company blowing past Meta, Tesla and edging out Amazon, Alphabet is what must sustain. Fundamentals that decelerate are when the exuberance tends to wear off, and that is right around Fall of 2024 as of now.

For EPS the growth decelerates in line with revenue growth:

• FY2024 ending in January is expected to report full year EPS of $12.40 for growth of 271.1%
• FY2025 has estimates of $21.36 EPS for growth of 72.32% — as stated, right now, this is front half weighted
• FY2026 has estimates of $26.54 EPS for growth of 24.24%

We broke our portfolio management rules of having a position above 10% allocation, and therefore, we have to take it seriously that both the top line and bottom line will decelerate from >200% to 40% over the span of six months. Most analysts are in agreement that a beat/raise is likely after hours tomorrow and perhaps for next quarter. What we are keeping an eye on is further out when Nvidia laps the strong quarters in October of this year. That is a long way off, but the market is forward-looking by about 9 months.

According to current estimates, there is no acceleration on the horizon through 2026. We think those estimates will ultimately be wrong especially once AI software ramps, but for now, this is the estimates investors are working with for pricing the stock.

Margins:

• Gross margin of 74.5% expected this quarter compares to GAAP GM of 63.3% in the year ago quarter. This equals $14.9 billion in gross profit. The adjusted gross margin is expected to be 75.5% this quarter.
• Operating margin of 58.7% is expected this quarter for operating profit of $11.7 billion. The adjusted operating margin is expected to be 64.5%.
• Last quarter, net margin was 51% for net profit of $9.25 billion and adjusted net margin was 55.3% for adjusted profit of $10.02 billion.

Cash Flow:

When we compare the world’s most valuable companies, Apple stands out for its cash. This is where Nvidia will have to improve to ultimately surpass Apple. During the hype cycle of AI software is where that is most likely to occur. 

Nvidia’s cash flow is still strong, yet it doesn’t hurt to compare it to other Mag 7 stocks:

Nvidia is the third strongest cash flow generator in the Mag 7, with an operating cash flow margin above 40%, but its smaller scale puts it in sixth place in terms of cash flow generation on a dollar basis. Nvidia’s $17.5 billion in TTM free cash flow pales in comparison to Microsoft’s and Alphabet’s nearly $70 billion – and while it’s not necessarily fair to compare companies in different tech verticals, Nvidia’s rapid ascent to a valuation above Alphabet and Amazon at some point will need to be reflected in the scope of its cash flows, especially when growth begins to decelerate.

Revenue Segments:

• Data center revenue last quarter was $14.5 billion, up 279% YoY. This compares to 31% growth in the year ago quarter. For this upcoming quarter, Nvidia is expected to report data center revenue of $16.9 billion for growth of 367%, which we outlined here along with a few different scenarios including how Nvidia can get to data center revenue of $101 billion in fiscal year 2025. Note that some of the data center revenue is also driven by networking for AI system with InfiniBand up 500% last quarter to $10 billion annualized run rate. We will update you more on networking after Nvidia’s report tomorrow and also when Marvell reports early March.
• Gaming revenue of $2.86 billion is up 81% YoY. This compares to a decline of (-23%) YoY in the year ago quarter.
• Pro Visualization was up 108% YoY for revenue of $416 million compared to a decline of (-65%) YoY in the year ago quarter.
• Automotive revenue of $261 million was up 4% YoY compared to 86% growth in the year ago quarter.

Additional Notes:

Analysts are Bullish

Bullish is the common theme heading into the report, given that Nvidia has raced from 13% growth in Q1 to 235% expected growth in Q4, and nothing describes the exuberance that accompanies this historic acceleration better than a handful of analyst estimates.

It’s within the data center that this bullishness is visible, as some analysts are expecting a nearly 20% beat on the Street’s $16.8 billion estimate, up to 60% higher than the Street by end of fiscal 2025.

Loop Capital is Nvidia’s largest bull heading into earnings, attaching a Street-high $1,200 price target on shares as the firm believes data center and overall revenue growth through FY 2026 will be meaningfully above the Street’s estimates. Loop is modeling a 17% beat in data center revenue to $19.6 billion, the highest on the Street, driving a 14% beat in total revenue to $23.1 billion. 

Loop is projecting the data center to reach a $100 billion annual run rate by fiscal Q2 2025, closing the year out with data center revenue of $117.5 billion, 41% higher than the Street’s consensus of $83 billion. Overall, Loop is modeling more than $132.3 billion in total revenue for Nvidia next year, 38% higher than consensus at $95.8 billion and representing 123% YoY growth, 65 percentage points above the Street.

It is entirely plausible that Nvidia’s growth continues to fly past expectations and mirror a scenario similar to what Loop is modeling, given the elevated levels of demand for its H100 combining with the launch of its faster H200 and B100 GPUs later this year.

KeyBanc sees that Nvidia’s AI capacity is well above the Street and can support data center revenues above $100 billion in calendar 2024, nearly 30% higher than what the Street is modeling. In that sense, there still may be room for another surprise in 2024.

UBS follows closely behind Loop with expectations for a similarly large data center beat in Q4 and impressive Q1 guide on strong demand for AI compute. Analysts are expecting Nvidia to beat on data center revenue by ~$2.5 billion to $3 billion, with their estimate at $19.5 billion for the segment and $23 billion for total revenue. UBS also believes that with “supply chain work,” Nvidia could guide to $25 billion to $26 billion in revenue for fiscal Q1, more than 16% above consensus estimates for $21.9 billion. 

BofA is more tame than Loop and UBS, calling for a modest 3-5% beat, or between $500 million to $1 billion above consensus for Q4’s report and Q1’s guide. This view for a beat and raise stems from supply gains offsetting impacts from China restrictions. However, BofA cautions that a beat of this size “’would pale vs. the 10%/22% beat/raise of prior quarters and perhaps disappoint some bulls,’ the more measured pace will also be seen as creating more fertile ground for continued growth.”

Meanwhile, going back to Q3’s report in November, analysts at Barclays said that the Nvidia's large Q3 beat “may not have cleared a very high hurdle,” and "didn't quite meet sky-high expectations" at "only" $2B ahead of consensus with margins at 75% and increasing into January. That commentary serves as a clear, yet somewhat brutal, reminder that even a $2 billion beat and raise had a muted response. 

Market Shifts in Anticipation of Growth Rate Changes

An interesting pattern has been playing out with Nvidia’s stock price over the past few years as its quarterly revenue growth rate has shifted.

Nvidia’s shares topped in November and December 2021, around 7 months before growth decelerated from the 50% range to just 3% growth in the July quarter. Shares bottomed in October 2022, 7 months in advance of revenues inflecting off a (21%) decline in the January quarter to a (13%) decline in the April quarter.

Current estimates are calling for a significant deceleration to just 38% growth in the October quarter, and if this pattern continues, then we are at the brink of setting a top above the $700 range as the market anticipates this deceleration. 

H200 and B100 to Launch in Q2 and Q4

Nvidia has an ambitious AI GPU roadmap, and is expected to release the next-gen H200 and B100 GPUs later this year, just over one year after releasing the H100.

The GPUs are expected to offer another leap in performance for AI training and inference, and the H200 is already in demand by the leading CSPs – AWS will be the first to deploy the new GPU, but Microsoft, Google and Oracle will also be deploying the chips. 

It’s easy to see why the cloud giants are eager to upgrade quickly — Nvidia says the H200 will boast reduced energy usage and thus a lower TCO, while the introduction of HBM3e memory will essentially supercharge the GPU’s performance. For GPT-3 175B, the H200 is expected to offer 1.4x to 1.9x faster LLM inference on the leading GPT and Llama models compared to the H100, and an 18x performance upgrade compared to the A100.

While it will be too soon to gauge what level of demand there is for the two new GPUs from a Q1 guide, a fiscal year guide could provide insight into whether demand for the H200 and B100 can match the H100, or if Nvidia will face initial supply constraints while ramping production of the two at the same time. Additionally, Nvidia will face competition this year from AMD’s MI300s.

A Note on China:

We detailed in our Q3 report the risks surrounding China given its importance to Nvidia as well as the export restrictions impacting Nvidia’s ability to sell the A100 and H100. Nvidia’s CFO said last quarter that “export controls will have a negative effect on our China business, and we do not have good visibility into the magnitude of that impact even over the long term.”

Any China commentary will be critical, given the $80 billion to $100 billion data center segment that may be impacted. Keybanc has the $101 billion estimate for the data center segment this year yet believes that $20 billion is dependent on China. Per our write-up: “Keybanc sees a $5 impact to Nvidia’s $25.62 EPS estimate, and up to a $20B impact to its data center segment with current estimates at $101B for the data center in FY2025.”

Valuation:

Fundamentally, Nvidia’s valuation is still quite cheap compared to historical benchmarks, given the sheer leverage and earnings power that the H100 is driving.

Shares are trading at a 91x PE and 32x forward PE ratio, and though it may look elevated, it’s not a range that Nvidia is unaccustomed to – shares traded between a 75x to 100x PE ratio for a majority of the time from the second half of 2020 to early 2022. However, Nvidia’s forward PE of 32x is where the valuation has room to run. Shares bottomed in October 2022 at a 34x forward PE with declining revenue and EPS, compared to today, where EPS is expected to grow at least 73% YoY to $21.36.

On a PS basis, Nvidia still looks reasonably valued, with more potential upside if it can surprise again in 2024. Shares are trading at a forward PS of just over 18x, around the same level it held through the second half of 2023 and a steep discount to the 32x forward PS it peaked at in late 2021. Prior to 2023, Nvidia last traded at around an 18x forward PS in July and August 2022, despite the challenging macro headwinds and declining revenue growth.

Conclusion:

Our process is such that we have no issues placing Nvidia at a lower allocation if needed, or increasing that allocation back to the #1 position if needed. We want to remain flexible while acknowledging two things – the first, is the company will eventually lap high comps and the sky-high growth will not sustain forever. Secondly, that this company is the defacto leader in the multi-generational investment opportunity of AI and is trading at a reasonable valuation.

It’s entirely plausible we get a beat/raise tomorrow and a beat/raise for the next quarter. What needs to be watched is the H2 estimates as they lap high comps of 200%+. The first graph above best illustrates this.

With that said, we are in the first, early powerful move for AIfirst, early powerful move for AI. We have two more powerful moves to go — AI software and AI at the edge, and then automotive will be the grand finale. Nvidia is a leader in both, and I’ve been quite clear that our stance is that AI software for Nvidia specifically will drive more revenue than AI accelerators. We are seeing early indication that Nvidia can and will compete with Big Tech on AI software and AI at the edge with the Chat with RTX application.

Stay tuned for our post-ER writeup to hit your inboxes tomorrow night.

Premium Members, you can look forward to a deep dive on a stock that is new to the IOF portfolio that we plan to accumulate this year – we called Meta the one that got away in our year end webinar. This stock is the runner-up and we think it has more room to go (fingers crossed). Look for this in your inboxes next week.

 Recommended Reading:

• Crypto and AI Opportunity: Real Vision Video Interview
• Cloudflare Q4: Key Metrics Accelerate
• Special Webinar Replay – February 1, 2024
• AMD Q4 2023 Earnings: 75% GPU Raise, Separating the Wheat from the Chaff

Memory plays a critical role in the world of AI, as faster and more powerful AI chips and servers will require increasing amounts of memory. The recent surge in generative AI and AI GPUs, spurred by the success of OpenAI’s ChatGPT and development of hundreds of other large language models, are forecast to bring about a new DRAM market, underpinned by high-bandwidth memory (HBM) and DDR5. Micron is uniquely positioned to benefit from this secular AI growth tailwind, leading the industry with the fastest, highest-capacity HBM, alongside other industry-leading products for AI applications.

This analysis will touch on how the memory market is recovering off one of the worst cyclical downturns it has faced recently, how NAND and DRAM are evolving, the massive shift ahead for HBM3, and how this translates to a multifaceted growth opportunity for Micron to tap into to push towards record revenues.

Please note, our plan is to patiently wait for the right entry. This is detailed below at the Conclusion.

Memory Market Recovery Begins in 2024, Accelerates in 2025

The memory market is coming off its worst cyclical downturn in 15 years, with late 2022 seeing memory revenues fall off a cliff, with challenges persisting through much of 2023. Heading into Q4, the market is showing multiple signs of bottoming – South Korean NAND flash exports returned to growth in September, NAND and DRAM pricing is expected to return to growth this quarter, and supply bit shipments are forecast to return to double-digit growth in 2024.

The downcycle starting in Q2/Q3 of 2022 exceeded the previous cycle in late 2018 in terms of its scale and duration – memory revenues declined approximately (-26%) from ~$44 billion in Q2 to nearly $33 billion in Q3. NAND revenues declined (-24.3%) to $13.7 billion, while DRAM fared slightly worse, with revenues declining (-28.9%) to $18.2 billion.

Q4 of last year saw another ~ (-27%) sequential decline to just over $24 billion in revenues. NAND revenues slipped around (-19.7%) sequentially to ~$11 billion, while DRAM revenues fell (-34.2%) sequentially to $12.3 billion. Essentially, DRAM revenues saw nearly a (-50%) decline in just two quarters, while the broader memory market declined just over (-45%).

Driving this rapid slowdown was a rapid deterioration in NAND and DRAM pricing – this was exacerbated by excess inventory at major manufacturers Samsung and SK Hynix leading to a fire sale at extremely low prices. NAND and DRAM both saw pricing decline more than (-20%) QoQ in Q3, and nearly (-30%) QoQ in Q4 2022. The previous year, from Q3 2021 through Q1 2023, NAND prices fell (-55%), and DRAM declined (-57%).

Source: Yole Intelligence

Pricing is expected to bottom out in Q3, with Q4 projected to see the first QoQ increase in both NAND and DRAM pricing since Q3 2021. That trend is set to continue through 2024, with Yole Intelligence forecasting NAND and DRAM prices to continue rising.

Source: Bloomberg

South Korean export data further supports the recovery story for NAND and DRAM.

NAND flash exports for September rose for the first time in a year, while DRAM exports registered the smallest decline. NAND flash exports increased +5.6% for the month, compared to an (-8.9%) decline in August, while DRAM exports fell (-24.6%). In October, chip exports from the country declined just (-3.1%) YoY to $8.9 billion in October, the smallest decline since August 2022, and another data point in support of memory’s recovery from the deep trough of late 2022 and early 2023.

A Broader Look at Memory’s Rebound

Pricing and export data both are signaling a return to growth for the memory market starting in Q4 and persisting through 2024 and 2025. Production cuts beginning in early 2023 are expected to lead to an undersupply of chips over the next four to eight quarters, and combined with steadily increasing NAND and DRAM prices, the memory market is projected to jump to record levels by 2025. cuts beginning in early 2023 are expected to lead to an undersupply of chips over the next four to eight quarters, and combined with steadily increasing NAND and DRAM prices, the memory market is projected to jump to record levels by 2025.

Overall, the memory market is projected to register a (-41%) YoY decline to approximately $84 billion in revenues in 2023, down from ~$144 billion in 2022. DRAM revenues are projected to fall (-47%) YoY from $79.7 billion in 2022 to ~$42 billion in 2023; NAND revenues are expected to fall (-37%) YoY from $58.7 billion to $37 billion this year.

2024 is forecast to see a sharp rebound in the market, with revenues rising approximately +55% YoY to more than $130 billion, according to Yole Intelligence. The increase in prices are driving the revenue jump with DRAM growing by “as much as 87 percent” in 2024 while NAND flash memory is projected to “bounce back to grow by about 60 percent,” according to Gartner.

2025 is projected to see the market reach record revenues of above $200 billion, or over +55% YoY for a second year straight, boosted by increased prices, undersupply – especially in DDR5 and other DRAM sub-segments — and AI mega-trends.

According to Lam Research, AI servers use 8X DRAM and 3X NAND compared to an enterprise class server.  For DRAM, that AI server and data center demand is expected to push the markets to new highs, from that ~$42B size in 2023 to ~$96B in 2028, according to Yole Intelligence.

Overview: HBM and DDR5’s Role in AI

High-bandwidth memory (HBM) – more specifically the next-gen HBM3/HBM3e – and DDR5 play a mission-critical role in AI, and such a role will lead the market to stunning growth: market leader SK Hynix projects the HBM market will grow at an 82% CAGR through 2027.

High bandwidth memory (HBM) offers higher bandwidth, capacity, performance, and lower power by vertically stacking up to twelve DRAM memory chips to shorten how far data has to travel, while also allowing for smaller form factors. Stacked memory chips are connected through something called “through silicon vias” or TSVs. HBM is increasingly being used to power machine learning, high performance data centers, and more recently, generative AI models. For a different perspective into HBM from an equipment leader, read our analysis on Lam Research here.here.

DDR5 DRAM, or double data rate 5, aimed to double bandwidth and data transfer speeds at a lower latency and power consumption than its predecessor, DDR4. DDR5 memory chips can be mounted on circuit boards to create memory modules, for use in servers or PCs. DDR5’s increased bandwidth allows for faster processing for memory-intensive applications, such as generative AI and training LLMs.

Demand for high-capacity memory is being driven by the sudden rise in generative AI and LLMs, which both require significant amounts of computing power and substantial amounts of DRAM to meet elevated performance requirements. SK Hynix’s head of DRAM marketing Park Myung-soo explained that “an AI server requires 500-gigabyte (GB) or larger High Bandwidth Memory (HBM) chips and at least 2-terabyte (TB) DDR5 chips.” As such, HBM3 and DDR5 are projected to see a rapid shift to become the dominant architectures over the next few years.

In 2022, HBM2e was the dominant HBM architecture, accounting for ~70% share of HBM shipments, with the emerging HBM3 taking just 8%. However, HBM3 will see a surge in demand in 2023 and 2024, rising to 39% share this year and approaching 60% share in 2024 to become the dominant architecture. This will be driven by the massive demand for Nvidia’s H100 GPU and AMD’s upcoming MI300, which are underpinned by HBM3.

As a result, HBM3 revenues are forecast to rise as much as +127% YoY to $8.9 billion, according to TrendForce, while SK Hynix estimated that this AI chip boom will lead to the HBM3 market expanding at an +82% CAGR through 2027.

Similar to HBM3, DDR5 is expected to quickly become the mainstream DDR architecture, driven by demand for faster compute for AI. DDR5 was expected to take more than 25% share of bit shipments in 2022, before rising to take more than 55% share in 2023. By 2026, DDR5 is projected to hold almost 95% share of bit shipments as the memory market completes its transition over to DDR5 from DDR4.

Source: Tom’s Hardware

Micron is well positioned to benefit from this massive shift to HBM3 and DDR5, as it is first-to-market with its 128GB capacity DDR5 RDIMMs (registered memory modules) and its eight-high 24GB HBM3e cube.

Background on Micron’s Positioning in HBM3 and DDR5

Throughout 2024, Micron is seeing “accelerating AI-driven opportunities for memory and storage across multiple market segments from the data center to the edge,” and it is rolling out industry-leading HBM and DDR5 products. Micron’s 8-high 24GB HBM3e cube began sampling in late July 2023, and it plans to sample its 12-high 36GB HBM3e cube in the first quarter of 2024 – among the first in the industry to reach the market.

Micron is investing heavily to capitalize on this HBM3/3e shift, with management noting last quarter that “assembly and test capex is projected to double year over year in fiscal 2024, predominantly driven by investments to support HBM3e production.”

Micron currently has deployed the industry’s fastest and highest capacity HBM3e on the market, supporting the most advanced AI training and inference. Micron’s HBM3e can deliver faster training times and more responsive queries for LLMs — it “increases performance per watt resulting in lower time to train LLMs such as GPT-4 and beyond.”

Micron’s HBM3e provides higher memory bandwidth that exceeds 1.2TB/s and 50% more memory capacity per 8-high 24GB cube, improving the accuracy and precision while training LLMs. Micron explains that this allows for up to 50% or more queries per day while reducing training times by 30%, thus lowering total cost of ownership (TCO).

TCO is an important factor for hyperscalers when evaluating equipment, especially GPUs, as it factors in not only the acquisition cost but also the costs associated with owning and operating the equipment over the hardware life cycle. Micron’s 24GB cube touts 2.5 times performance per watt improvement over previous generations; this increased power efficiency generates “tangible cost savings” for AI data centers. Micron explains that for “an installation of 10 million GPUs, every five watts of power savings per HBM cube is estimated to save operational expenses of up to $550 million over five years.”

In terms of DDR5, Micron says it currently sits as the market leader, with the most market share in the early innings of this DDR5 shift, underpinned by the memory industry’s most developed DDR5 ecosystem. Micron has launched its high-capacity 96GB DIMMs, and at the beginning of November, Micron launched the first-to-market 128GB DIMM based on its 1β technology, which it says “delivers the fastest speed and lowest latency” of any DDR5 DIMM available.

Micron explains that the 128GB DIMM offers more than 24% improved energy efficiency, as well as 16% improved latency, which is crucial for “memory-bound workloads such as generative AI, in-memory

databases, and real-time data analytics, where high-capacity is needed, and prompt response times are critical for real-time inference.” Micron adds that the 128GB DIMM “delivers up to 28% faster performance for AI training” on models such as Meta’s Llama 2-70B.

Customers are seeming optimistic about the benefits that the 128GB DIMM offers – AMD SVP Dan McNamara said that “as AMD advances compute with our next-gen EPYC processors, Micron’s 128GB RDIMMs will likely become one of the main memory options to deliver high-capacity and bandwidth per core capabilities to address the demands of memory-intensive applications.” Intel VP Dr. Dimitrios Ziakas echoed that sentiment, saying, “Intel is evaluating this 32Gb memory offering for key DDR5 server platforms based on the resulting total cost of ownership benefits to cloud, AI and enterprise customers.”

Competition Remains Stiff for Micron

While Micron claims it is first-to-market with HBM3e and holds the most market share in DDR5, competition remains stiff, as Micron is competing against two heavyweights who control more than 85% of the market – Samsung and SK Hynix. The two South Korean firms are rapidly advancing HBM3 development, with SK Hynix already firmly established in the market as it was the preferred HBM3 supplier for Nvidia’s highly popular H100 GPU.

SK Hynix unveiled its 1TB bandwidth HBM3e memory in late Q2 this year, with mass production set to start in early 2024. SK Hynix is also reportedly eyeing development of its next-gen HBM4 cube with a plan to introduce that product to market in 2026.

Samsung is currently mass producing its 12-high 24GB HBM3 cube, ‘Icebolt’, and is sampling its HBM3e cube ‘Shinebolt’ to prospective customers. While coming to market later than Micron and SK Hynix, Shinebolt is rumored to compete with Micron on performance, with both offering 1.2 TB/s bandwidth. Samsung is also expected to unveil its fifth-gen HBM3e cube, named ‘Snowbolt’, by the end of the year, followed by a sixth-gen HBM cube next year.

Nvidia, AMD Battling on Memory

HBM3 and HBM3e are becoming the next battleground for memory chip manufacturers as well as AI chip developers, especially Nvidia and AMD, who are pushing the boundaries with the amount of memory bandwidth in each GPU.

AMD’s competing GPUs, the MI300 series, substantially boosted memory and bandwidth relative to the H100, utilizing Samsung’s HBM3. The MI300A is shipping with 128GB HBM3 memory while the MI300x ships with 192GB memory and 5.2 TB/s of bandwidth – that’s 1.6x more bandwidth and 2.4x more HBM3 density than Nvidia’s H100.

Nvidia is rapidly moving forward with its GPU roadmap, as it aims to launch its next-gen H200 and B100 GPUs next year followed by the X100 GPU in 2025 – each GPU will accelerate AI inference times along an exponential curve, thus creating a need for more memory and more bandwidth.

Source: Nvidia

Nvidia’s A100 shipped in two different versions with either  40GB or 80GB HBM2e memory, with the 40GB offering 1.55TB/s of bandwidth and the 80GB offering 2TB/s bandwidth, the industry’s fastest at the time in 2021.

Nvidia then upgraded from HBM2e to HBM3 DRAM, tapping SK Hynix as the supplier for its H100 GPU for 1.6X the bandwidth. Nvidia’s upcoming H200 GPU, set to launch in early 2024 as the industry’s first HBM3e-powered GPU, is expected to ship with another 1.5x bandwidth boost relative to the H100 with nearly 1.8x the memory.

It is rumored by some sources that the H200 is shipping with Micron’s HBM3e, instead of SK Hynix. rumored by some sources that the H200 is shipping with Micron’s HBM3e, instead of SK Hynix. Micron reportedly sampled its 24GB HBM3e memory with Nvidia at the end of July, with SK Hynix following in mid-August and Samsung in early October. According to sources in South Korea, Nvidia remains engaged with SK Hynix for the H200.

This raises an important point about competition in this AI chip and memory race: if Nvidia is switching this quickly from one supplier to the next based on time to market, this raises the risk that Samsung or SK Hynix could be first to market with a superior HBM4 product and take share away from Micron, especially if they undercut Micron on price.

AI Will Increase Secular Growth Opportunity for Micron

The recent surge in AI and data center growth fueled by Nvidia is expected to translate into an interesting shift in revenue mix for Micron: it foresees exposure to the more cyclical and seasonal PC and mobile end market declining from 55% share of revenue in FY21 to 38% share in FY25.

Data center and graphics are forecast to rise from 30% share of revenue in FY21 to 42% share by FY25, with AI and machine learning driving such growth; the projected surge in the HBM3/HBM3e market supports this shift. In data center, Micron is expecting NAND GB shipped per server to increase 3x and DRAM 2x by 2025, as AI servers require significantly more memory than traditional servers.

Automotive and industrial are projected to rise from 15% share to 20% share, as both end markets exhibit much faster growth rates than PC and mobile due to the rise of electric vehicles, industrial robotics, and other emerging trends which require a higher semiconductor content per unit.

This revenue mix shift is underpinned by long-term agreements, at ~75% of revenue in CY22. This offers multiple benefits: more visibility into forward revenues, less exposure to cyclical pricing trends in NAND/DRAM with pricing locked in for the contract duration, reduced impacts from supply and demand imbalances, and ultimately more stable margins.

Financials Sharply Improving

The swift decline in the broader memory market over the past eight quarters has had a significant effect on Micron’s financials. Revenues plummeted, falling (-49.5%) YoY in FY23 from $30.76 billion to $15.54 billion. Operating margin also shifted deep into the red, with Micron posting a (-37.0%) operating margin, compared to a 31.5% operating margin in FY22. Micron reported a net loss of (-$5.83 billion), or ($5.53) per share, compared to net income of $8.69 billion, or $7.75 per share, in FY22. This was the sharpest decline for revenues, operating income, and net income in Micron’s history.

Fiscal Q4 (ending Aug 31) showed initial signs of a recovery:

• NAND revenues increased 19% sequentially to $1.2 billion, bit shipments rose >40%
• DRAM revenues increased 3% sequentially to $2.8 billion
• Total revenue increased 7% sequentially to $4.01 billion
• GAAP net loss improved 25% sequentially to ($1.43 billion)
• Operating cash flow improved 938% sequentially to $249 million, but is much lower compared to $3.78 billion in the year ago quarter

Fiscal Q1’s guide was boosted at the end of November:

• Revenue is projected to be $4.7 billion, compared to a prior view for $4.4 billion +/- $200 million. This would represent YoY growth of 14.9%.
• Gross margin is expected to be (0.5%) to 0%, compared to the prior view of (4.0%) +/- 2.0%

The surge in the HBM3 market, positive outlooks for a NAND and DRAM pricing recovery through 2024 and into 2025, and a surge in AI and data center demand are expected to fuel a rapid recovery for Micron’s top and bottom line.

Moving forward through FY24 (Sept. 2023-24, this reacceleration in NAND and DRAM, buoyed by increasing pricing, is expected to send Micron’s revenue on an eight-quarter streak with more than +20% growth – even as high as +65% as it laps easy comps. On a dollar basis, revenues are forecast to rise from $4.01 billion in Q4 of fiscal 2023 to $8.71 billion in Q1 FY26.

Essentially, Micron is on track to potentially reach record revenues just over eight quarters after that massive slump. However, EPS is forecast to be below levels seen in FY18 and FY22 – estimates peak at $2.05 in Q4 FY25, compared to $3.53 in Q4 FY18 and $2.59 in Q3 FY22. What this means is that revenues are being propelled higher by this shift to HBM3/3e (as it exhibits much higher ASPs relative to typical DRAM memory), but margins are having a tougher time recovering as rapidly due to the deep trough that NAND and DRAM prices must rebound from.

Operating cash flow is also expected to rebound quickly after plunging alongside revenues and EPS in FY23. OCF margin is estimated to rebound from 10% to more than 40% by FY25, with Micron projected to generate upwards of $13.2 billion in operating cash flow, compared to $1.56 billion in FY23.

Micron has applied for CHIPS Act funding for its New York and Idaho fabs, saying that federal funding and tax incentives were needed to develop both facilities. Micron is investing up to $115 billion over the next 20 years to build out its US production base, in an effort to boost the US’ share of production from 2% to 15% and diversify away from East Asia – Micron’s current high-end chip production is more concentrated in Japan and Taiwan. Given that construction isn’t set to begin until late next year with production commencing as early as 2026, any margin benefits from the CHIPS Act are unlikely to be recognized over the short and medium term.

Risks

China presents a real risk to Micron as it does to much of the semiconductor industry. Micron is generating nearly one-quarter of its revenues from China, and CEO Sanjay Mehrotra recently told CNBC that “about half that revenue is at risk.” Micron was the first American chip company targeted by China with a partial ban earlier this year, and the company is working to improve relations with the nation, though there is no guarantee that such a ban will be lifted.

However, there are risks to this recovery, in that it may not unfold as smoothly as projections picture. Quarterly revenues have been variable over the past few fiscal years, with multiple sequential declines present along the growth trend, so there is a risk that current projections calling for sequential growth through Q1 FY26 do not account for some of that lumpiness.

In addition, there is also a tail-end risk in that NAND and DRAM pricing does not exhibit consecutive sequential growth through 2024. This could be exaggerated if NAND pricing slips back to sequential declines in Q2, as it is forecast to see just +3-4% sequential growth in that quarter. Pricing has shown to be volatile in the past, and there is no guarantee that pricing will rebound smoothly and steadily. Should some sequential declines appear in NAND and DRAM pricing in 2024, this would likely weigh on both margins and EPS.

Valuation

Micron’s fundamental backdrop is projected to see rapid top-line growth and a sharp bottom-line improvement on the backs of surging HBM3/3e demand; however, semis have rallied this year due to Nvidia leading the historic Nasdaq rally in the first six months of 2023. This has resulted in a rising of all boats, as many semis are trading far above historical multiples. The product of strong gains in the broader semiconductor industry has pulled Micron’s shares higher: the iShares Semiconductor ETF has gained +48.7% YTD, significantly outperforming the S&P 500’s +18.7% return.

Micron currently trades at a 5.43x PS ratio, its highest level in more than 20 years, in part due to FY23’s (49.5%) YoY decline in revenues while shares have gained +54.5% YTD. Even with +34.5% estimated revenue growth in FY24 to $20.9 billion, Micron still trades at a 4.05x 1-year forward PS ratio, far above its 5-year median PS ratio of 2.75x.

Micron’s forward EV/EBITDA multiple of ~13.0x also is elevated, at nearly double its 5-year median multiple. This accounts for the expected top and bottom line recovery for next year, but it will take more than a few quarters for this top-line growth to translate to a strong recovery in EBITDA and regression to the mean.

Conclusion

As Beth stated in Nvidia’s Q3 earnings preview, HBM3e is rapidly making its presence known, with Nvidia’s upcoming H200 GPU to be the first with HBM3e memory, rumored to come from Micron (still needs confirmation, but is an exciting possibility should it come to fruition).

Nvidia’s rapid GPU upgrade roadmap in response to AMD’s MI300X is a testament to the fast-moving nature of not just the AI GPU market, but also memory – HBM3e is coming to market in 2024, but may quickly be replaced by HBM4 in 2025 and future iterations of HBM memory beyond 2025 to 2026. 

It is a highly concentrated market, dominated by Samsung and SK Hynix, though Micron remains an important player as it moves ahead with industry-leading first-to-market HBM3e and DDR5 solutions. Micron looks well positioned to capitalize on this AI mega-trend, with revenues from both solutions contributing in 2024. On a broader scale, AI and data center are set to transform Micron’s revenue mix to stronger and more secular end markets from its historically cyclical and seasonal concentration in PCs and smartphones — Micron estimates data center and AI to rise from 30% of revenue in FY21 to 42% in FY25, while smartphone and PC’s 55% share is estimated to decrease to just 38%.

However, Micron is valued at elevated multiples, when memory stocks typically would be on a deep discount given the steep downcycle, yet SOXX returns are > QQQ returns. This has lifted the tide of all boats, and memory stocks such as Micron are not trading where they’d normally trade, which makes a near-term buy less likely for the I/O Fund until we see a pullback. This is where the I/O Fund is unique, not only do we strive to be early in our research, such as to the importance of memory for the next generation of AI accelerators, but we are also careful with our timing.

Technical Analysis

The long-term pattern best fits as a large degree diagonal pattern. This is a 5 wave uptrend that is characterized with large corrective swings. If accurate, we are in the middle of the 4th wave correction, which would target the $48 – $35 region before completing.

If we zoom into the 2021 top to now, we can get a better look at the risk parameters that would either confirm or invalidate this setup. For one, the price action from the 2021 high best fits this pattern. Note how we only have a clear 3 wave drop into the October, 2022, low. This is followed by a messy and overlapping uptrend into the recent high. More times than not, when the following bounce after a 3 wave drop is a messy and overlapping move, it signals a bounce in a larger corrective pattern.

If we see a break down below the $62-$58 support region, then this pattern will be confirmed, as we establish lower targets to buy this stock. On the other hand, in order to invalidate the risk present in the current price structure, we need to see price break above the $84-$91 resistance zone. The higher we go into this region, the more likely that we will see a continuation of the larger uptrend.

I/O Fund Equity Analyst, Damien Robbins, contributed to this analysis. I/O Fund Portfolio Manager, Knox Ridley, contributed to this analysis.

Recommended Reading:

• Memory and PC Stocks Review
• Marvell Q3 Earnings: The Market Wants More on AI
• Marvell Q3: AI-Driven Rebound on the Books, Bottom Line in Focus
• Nvidia Fiscal Q3 Earnings: The China Impact
• 2024 Trend: Memory and PC Rebound

Nvidia’s Fiscal Q3 earnings report was spectacular on all accounts. The data center growth for this quarter was bonkers againbonkers again with growth of 279% year-over-year. The FQ4 guide implies data center growth that will accelerate to roughly 370% next quarter.

Overall revenue beat by $2 billion this quarter for $18.12B in revenue, up 206%. For next quarter, the guide beat by another $2 billion for guidance of $20B compared to $17.9B expected. 

EPS of $4.02 compares to $3.39 expected. Gross margin grew to 74% compared to 71.5% expected.

Yet, the stock is down 1.5% after hours. In our pre-earnings write-up that was published this morning in our free newsletter, I had stated: “The topping-out scenario is that Nvidia’s buying is exhausted, and there isn’t one fundamental analyst on earth that can help investors figure out when this will happen.” I explained this is best left to technical analysts as there are many forces which weigh on a stock. It was unlikely Nvidia missed due to the CFO’s visibility on supply, yet the win-win scenario is that we don’t need the market to continue to reward Nvidia. It already has rewarded Nvidia, and if the market is getting tired of Nvidia’s exceptional results, then we will simply take gains and buy again lower.

Essentially, what we are seeing after hours has nothing to do with the company’s financials. When buyers become exhausted, it means a story is well-known. It’s not logical, it’s merely what makes a market.

The flaw in Nvidia’s report is the loss of China revenue. It can be deceiving because demand is so high, that Nvidia will absorb those losses in the upcoming quarter. However, there are implications in the medium-term, which we had also written about in our pre-earnings report.

The long-term thesis is very much intact, which is that Nvidia is on its way to become the world’s most valuable company someday. Data center GPUs are only part of the story. Automotive has the potential to exceed data center GPUs, and there’s also software which we covered here.

In the near-term, Nvidia investors should keep an eye on the broader semiconductor sector, which looks weak, and there’s a chance the China impact drags on FY2025/FY2026 estimates until we get a new fiscal year guide next quarter. I also touch base on a few positives that are important to keep an eye on.

Revenue and EPS:

Nvidia reported revenue of $18.1B, up 206% Y/Y and well above consensus of $16.1B and above guidance of $16B. However, the magnitude of the revenue beat of 12% was smaller than the 22% beat in the July quarter. Non-GAAP EPS was $4.02, well above consensus of $3.39.

Revenue Segments:

• Data Center revenue of $14.5B, up 279% YoY and up 41% QoQ
• Gaming revenue of $2.9B, up 81% YoY and up 15% QoQ
• Pro Visualization revenue of $416M, up 108% YoY and up 10% QoQ
• Automotive revenue of $261M, up 4% YoY and up 3% QoQ
• OEM & Other Revenue of $73M, flattish YoY and up 11% Q/Q

Nvidia provided revenue guidance of $20B +/- + 2% above consensus of $17.9B with adjusted GM guidance of 75.5% and Non-GAAP Operating Margin guidance of 64.5%.

More on Data Center Segment:

Our pre-earnings report highlighted the release of the H200. Major cloud players such as AWS, Google Cloud, Microsoft Azure, and Oracle cloud will be among the first CSPs to offer H200 inferences starting in Q2 of 2024. The H200 is likely to come with a higher ASP than the H100 due to HBM3e memory. The H100 has an ASP in the $30,000 to $40,000 range. The higher ASP may not contribute to margins necessarily, as HBM3e is costly.

At $40,000 per H100, that equals $29B in H100 sales alone, and when you add the A100 and other data center sales at a current run rate of $15B, the Data Center segment could report total revenue of $44B in FY24 (CY23). When you equal this out across the upcoming quarters, it looks something like this based on our estimates and Piper Sandler estimates.

Nvidia is expected to report approximately $16.5B in revenue for the January quarter. Keybanc has data center revenue at $101 billion for next year. If we assume China is $20 billion of this (and worst case, doesn’t get absorbed) then it will look something like this:

Scenario 1:

Q1 FY25: $18B

Q2 FY25: $19.5B

Q3 FY25: $21B

Q4 FY25: $23B

However, it’s likely the China revenue does get absorbed even if analysts are forced to revise estimates for now. This means that estimates may go down this quarter, and then be revised up again when management discusses the fiscal year guide. If so, it would look more like this:

Scenario 2:

Q1 FY25: $20B

Q2 FY25: $24B

Q3 FY25: $27B

Q4 FY25: $30B

That’s based on Keybanc’s fairly optimistic estimate of over $100B next year in data center revenue. Here are data center revenue numbers that are more conservative from Piper Sandler. Due to the QoQ growth in this model, next quarter’s fiscal year guide is paramount for us Nvidia bulls.

Scenario 3:

My opinion is that Scenario 1 is a safe assumption as it combines continued growth in the data center with some China impact.

Margins:

Gross margin of 74% beat guidance of 71.5%. As stated in our pre-ER write-up, these are historic margins for Nvidia.

The company reported an operating margin of 57.5% for income of $10.4 billion. The adjusted operating margin of 63.8% compares to a margin of 26.4% last quarter.

Net income of $9.2 billion represents a margin of 51% compared to 11.5% net margin in the year ago quarter. This is a combination of data center strength and being at the cyclical trough last year for gaming.

Cash:

Cash flow margins are the best in the Mag 7 at 40.5% operating cash flow this quarter and 38.9% in free cash flow margin. Meta has the second best FCF margin at 34.7% followed by Apple at 29.7%.

Nvidia had $18.3B in cash and marketable securities, up from $16.0B last quarter and debt of $9.7B in-line with the July quarter of $9.7B.

The company utilized cash of $3.91 billion towards shareholder returns, including $3.81 billion in share repurchases and $99 million in cash dividends. Last quarter, an additional $25 billion was authorized for share repurchases.

Earnings Call:

The China Impact:

We had written the following in our pre-earnings report:

“The Red Scare:

What’s not to be forgotten in the excitement of the product road map is China, which has been the predominant risk for semiconductor stocks dating back to 2018. Last year, the government restricted Nvidia from selling its two most powerful chips to China, the A100 and H100. To circumvent these restrictions, Nvidia designed slightly less powerful chips called the A800 and H800. As reported by Reuters, the H800 has as much computing power as the H100 in certain settings. For the United States, these chips are important to block as they strengthen China’s military.

Last month, the U.S. Department of Commerce announced updated rules focuses on computing performance by removing the bandwidth parameter and focusing exclusively on how powerful a chip is, as well as performance density, which will prevent companies from working loopholes. According to an official who spoke to Reuters, “the U.S. will require companies to notify the government about semiconductors whose performance is just below the guidelines before they are shipped to China.”

Although this is a medium-term issue for Nvidia, analysts believe the demand is high enough today that the company shouldn’t have any issues absorbing the 20% to 25% loss in its data center segment from tighter export restrictions to China. Looking further out for FY2025, Keybanc sees a $5 impact to Nvidia’s $25.62 EPS estimate, and up to a $20B impact to its data center segment with current estimates at $101B for the data center in FY2025.

Eventually, demand may settle – especially as more competitors step up – and investors should pencil-in losing China revenue as a risk that is materializing now, with the revenue impact likely to be felt in FY2025.”

It’s tempting to shrug off the loss of revenue given Nvidia beat/raised next quarter, which is the quarter when 20% to 25% of revenue from China and other restricted countries will be cut off.  However, the Street is likely to be cautious tomorrow because FQ4 will be seen as an outlier where demand can absorb the 20% to 25%. Basically, the outsized demand will be transitory whereas the U.S. Department of Commerce is cutting off 20% to 25% permanently. There was some talk about Nvidia serving these countries with a less powerful chip, but the restrictions are blacklisting Nvidia’s AI chips (specifically) so this workaround won’t be an easy feat.

By the time Nvidia comes up with a workaround, even if it’s acceptable, those countries will have designed their own domestic silicon. Even if this eventually does get absorbed, analysts will likely revised down their estimates for a few quarters out in FY2025 or next fiscal year FY2026. This may, in turn, impact Nvidia’s valuation. Per the CFO: “The export controls will have a negative effect on our China business, and we do not have good visibility into the magnitude of that impact even over the long term.”

This does not derail Nvidia’s thesis by any means and the timing could not have been better with the restrictions happening during a period of outsized demand. As pointed out on the call, Nvidia will be tapped by many countries that are not blacklisted into the foreseeable future: “National investment in compute capacity is a new economic imperative, and serving the sovereign AI infrastructure market represents a multibillion-dollar opportunity over the next few years.”

InfiniBand up 500% YoY:

We covered InfiniBand a few years back when our site covered the Mellanox acquisition. Mellanox was an important acquisition as it helped Nvidia align its architecture with speed by supporting Virtual Protocol Interconnect (VPI), which allows the ubiquitous Ethernet to provide bandwidth as cheap as possible, and InfiniBand to deliver higher throughput and fewer bottlenecks during high loads. Today, this acquisition is paying off.

Per the opening remarks: “Networking now exceeds a $10 billion annualized revenue run-rate. Strong growth was driven by exceptional demand for InfiniBand, which grew fivefold year-on-year […] Azure uses over 29,000 miles of InfiniBand tabling, enough to circle the globe.”

InfiniBand growing five-fold exceeds overall data center revenue given the $15B in total data center revenue last fiscal year is expected to grow 200% to $45 billion at the exit of this fiscal year. Networking revenue tripled and data center compute grew four-fold.

The discussion on the call is that companies are standardizing with InfiniBand as the “computing fabric” increases the effectiveness of AI infrastructure by 20% to 30%. InfiniBand is nearly ubiquitous in supercomputing and is becoming popular with AI/Big Data applications on a large scale for high performance clusters. The benefits of the software defined fabric is that it’s low latency, high bandwidth and low management cost.

Recurring Software Revenue at $1 Billion:

Going off what we know, recuring software revenue may have doubled over the past few quarters CFO had stated: “hundreds of millions of dollars annually” and it’s now being stated the standalone software business will be worth $1 billion next quarter: “We are on track to exit the year at an annualized revenue run-rate of $1 billion for our recurring software support and services offerings.”

Keep an eye on this as it’s likely to be the leading story over the next few years – especially as automotive ramps.

AI Factories:

This was probably the most important question in terms of Nvidia’s growth potential. There’s nothing revelatory being said, per se, but it’s nice to hear some of the bigger picture repeated.

Question: “Because when I just look at the trajectory of your Data Center, it will be close to nearly 30% of all the spending in Data Center next year. So what metrics are you keeping an eye on to inform you that you can continue to grow? Just where are we in the adoption curve of your products into the generative AI market?” -Vivek Arya, Bank of America

Answer: “Generative AI is the largest TAM expansion of software and hardware that we've seen in several decades. At the core of it, what's really exciting is that what was largely a retrieval-based computing approach – almost everything that you do is retrieved off of storage somewhere – has been augmented now, added with a generative method. And its changed almost everything. […]

And one of the areas that is really impactful is the software industry, which is about $1 trillion or so, has been building tools that are manually used over the last couple decades. And now, there's a whole new segment of software called co-pilots and assistants. Instead of manually used, these tools will have co- pilots to help you use it, and so instead of licensing software – we will continue to do that of course, but we will also hire co-pilots and assistants to help us use the software. […]

But there's a new class of data centers, and this new class of data centers, unlike the data centers of the past, where you have a lot of applications running used by a great many people that are different tenants that are using the same infrastructure, and that data center stores a lot of files. These new data centers are very few applications, if not one application, used by basically one tenant, and it processes data. It trains models, and it generates tokens. It generates AI. And we call these new data centers AI factories.”

Translation: If you separate AI from traditional data centers (and where data centers are headed), then Nvidia represents far more than 30%.

Conclusion:

We are tracking TSM, semiconductor indexes, and Nvidia’s chart for signs of exhaustion as outlined here. We are seeking a win-win scenario where we can lock-in gains, and then use that cash to buy Nvidia again at lower levels. As stated, Nvidia’s thesis is firmly intact. Rather, the issue is the market is seeing very narrow leadership and Nvidia is the defacto leader within that narrow leadership. The saying in Wall Street is that pigs get slaughtered. That’s a rough way of saying 200% gains YTD should be approached carefully as the goal is to make real money, not paper money. Of course, 200% will be nothing by the time we are done with this position. But for this year, it’s good enough. 

 Recommended Reading:

• 2024 Trend: Memory and PC Rebound
• Big Tech companies continue to invest in AI
• Apple Q4: iPhone Revenue Accelerates while Services Shine
• Supermicro Fiscal Q1: “Conservative” Guide
• AMD Q3 Earnings: $2B in GPU Revenue for 2024

Initially, AMD sold off (-5%) based on the missed Q4 revenue guide. After hours, the stock recouped and ended up flat once management clarified the miss is not from the data center segment. Rather, it was confirmed the Q4 data center will grow 50% QoQ, and perhaps most importantly, management guided for $2B in GPU sales in 2024 from AI rather than supercomputing. This was the moment the price action reversed, and is a positive as it translates to broader hyperscaler customers driving the $2B rather than the Department of Energy’s supercomputer El Capitan.

We had stated in our pre-earnings write-up: “[The 50% revenue growth in Q4] is exciting, yet the market will want to see more balanced, commercial demand beyond the government-owned supercomputer.” AMD’s management went out of their way to provide additional color on hyperscalers driving the 2024 revenue, and I’ve included that part of the transcript below.

Roughly speaking, the data center is a $6 billion segment for AMD after 10 years of Lisa Su being the CEO (she became CEO in 2014). This means within one or two quarters of the GPUs shipping; sales will be the equivalent to 30% of today’s data center revenue right out the gate. Lisa Su stated the MI300 will be the fastest product they’ve had reach $1 billion in sales – which is saying a lot because the EPYC CPUs are popular.

Per the opening comments: “Based on the rapid progress we are making with our AI road map execution and purchase commitments from cloud customers, we now expect Data Center GPU revenue to be approximately $400 million in the fourth quarter and exceed $2 billion in 2024 as revenue ramps throughout the year. This growth would make MI300 the fastest product to ramp to $1 billion in sales in AMD history.”

Notably, it’s out of character for this management team to discuss forward revenue beyond one quarter. I find this break in style interesting, and I also take this to mean the $2 billion is a baseline. We will see, but that’s how I interpret a very early guide that would normally come in February. 

What is a bit unfortunate is that the gaming segment and embedded is weak, and these two segments overshadow AMD continuing to take market share on the CPU data center, and the company’s highly anticipated answer to Nvidia’s H100s. Per the CFO’s opening remarks: “In the fourth quarter, we expect to benefit from strong Data Center and Client momentum, driven by MI300 AI accelerated ramp and the strength of our high-performance leadership Zen 4 family of products despite lower sales in the Gaming segment and additional softening of demand in the embedded market.”

What is quite fortunate, however, is that some of AMD’s AI story is undercover, which is the PC and mobile market. There’s a bonus waiting for AMD investors as the AI story plays out, and while data center GPUs take up all of the attention, there is another powerful AI trend that is silently building strength in the background. I mentioned this in a previous write-up: “When discussing AMD’s AI opportunity, it is vitally important that we not lose sight of the opportunity AMD will have to expand its AI portfolio to the Client Segment.” This has already begun with the Ryzen AI on-chip accelerator contributing to the rebound in the client segment this quarter.

Revenue and EPS:

Revenue and EPS for Q3 came in a tad higher than expected at 4% growth for revenue of $5.8 billion. The company reported EPS of $0.70 versus $0.68 expected.

The guide for next quarter was problematic as it was a miss with management guiding for $6.1B in revenue versus $6.39B expected. However, the miss is not coming from the data center, and the market is digesting whether AMD is a “buy” given two of its segments are quite weak (gaming and embedded).

Margins:

Overall, the company came in as expected on margins. Gross margin came in at 47% which is a positive. On the call, it was mentioned that as the GPUs ramp, the revenue will be accretive to gross margin.

• Adjusted gross margin of 51% was in line with management guidance for adjusted gross profits of $2.963B
• Operating margin of 4% for operating profit of $224M is an improvement over the past few quarters of single digit negative GAAP operating margin. Per the CFO during those quarters, GAAP OM would improve with the Client segment recovery.
• Adjusted OM came in as expected at 22% for Adj operating profits of $1.27B
• GAAP net income was $299M with a margin of 5%

The CFO said the following about continued margin expansion:

“And in Q3, we saw very significant improvement with our client segment gross margin. I think going forward, the pace of Client segment improvement will moderate, but it will continue to drive incremental gross margin improvement in Client segment […] I think going forward, it's really mix, primarily mix, is driving our gross margin, but we feel pretty good about second half next year when we can expand the Data Center significantly and especially Embedded segment start to recover, we should be able to drive more meaningful gross margin improvement in second half.”

Cash Flow:

Operating cash flow has improved from last quarter although is still quite a bit lower than AMD’s peak in 2022. Operating cash flow of $421M is up from $379M last quarter, but is down from $965M in the year ago quarter. This should continue to improve with the Client segment bottoming.

Free cash flow of $297M is similar – an improvement from last quarter at $254M yet down from $842M in the year ago quarter.

The company has $5.8 billion in cash and short-term investments with $2.467B in debt. The company repurchased $511 million in shares.

Revenue Segments:

Data Center:

Data center came in as expected for Q3 at $1.6 billion in revenue, driven by the 4th Gen EPYC CPUs which offset a decline in adaptive SoC products. Per the opening remarks: “EPYC CPU revenue grew by a strong double-digit percentage sequentially” and also “we gained server CPU revenue share in the quarter as fourth-gen EPYC CPU revenue grew more than 50% sequentially, crossing over to represent a majority of our server processor revenue and unit shipments.”

Management has confirmed Q4 will be 50% sequential growth, or $2.4 billion in revenue. Of this, $400 million is from GPUs.

Client Segment:

The Client segment looks to have bottomed in a big way. Revenue of $1.5B was up 42% year-over-year and was up 46% QoQ. A full recovery would be in the $2.5B range. AMD released the Ryzen 7000 Series, which has helped CPU sales. Management also indicated that their partnership with Microsoft is a catalyst on the horizon as the Ryzen 7000 Series will power the biggest advancement in Windows in over 20 years.

Per management’s opening remarks: “Looking forward, we are executing on a multiyear Ryzen AI road map to deliver leadership compute capabilities built on top of Microsoft's Windows software ecosystem to enable the new generation of AI PCs that will fundamentally redefine the computing experience over the coming years.”

The Client Segment is expected to be strong next quarter, as well.

Gaming:

Gaming revenue of $1.5B was down (-8%) YoY due to a decline in semi-custom revenue but was offset by AMD Radeon GPUs. This segment is expected to decline next quarter due to a steep console cycle.

Embedded:

This segment is weighing on AMD as it’s lapping very high comps. As stated in our Pre-ER writeup, embedded is lapping a quarter with 1868% growth. Per the opening remarks: “Looking ahead, based on our current visibility, we expect Embedded segment revenue to decline sequentially as customers continue working through elevated inventory levels through the first half of 2024.”

The weakness in the gaming and embedded segments is expected to last (and perhaps worsen) through at least Q1. Per the CEO:

“And then from an Embedded and Gaming standpoint, we would say Embedded, think about it down similar levels sort of in the teens compared to sort of Q3 was down in the teens and Q4 will be down in the teens. And then Gaming, from a console standpoint, we do expect that to be down a bit more than that. And then as we go into Q1, again without being — there are lots of things that need to happen. We would expect that both gaming and embedded would be down into Q1 as well and sort of the other comments would be more around seasonality.”

Earnings Call:

It was key that management provide color on this call regarding the MI300 demand. The first question was centered the hyperscalers and how the MI300 is being used:

Toshiya Hari:

[…] My first one is on the Data Center GPU business. You talked about '24 revenue potentially exceeding $2 billion. I was hoping you could provide a little bit more color. What percentage of this is AI versus supercomputing or other applications?

Lisa Su:

Your question as to how the revenue evolves, so the way to think about it is, in the fourth quarter, we said revenue would be approximately $400 million, and that's mostly HPC with some — the start of our AI ramp. And then as we go into the first quarter, we actually expect revenue to be approximately similar in that $400 million range. And that will be mostly AI so with a very small piece being HPC. And as we go through 2024, we would expect revenue to continue to ramp quarterly, and again, it will be mostly AI.

The second question is when the stock price recovered after hours because this is when management made it clear the Q4 miss was not data center related. AMD had previously guided for 50% sequential data center growth (we covered this in great detail here) and this was reiterated:

Aaron Rakers:

[…] And how has that $400 million evolved underneath that? Has that — was it $300 million now going to $400 million? Just how has that changed over the course of the last quarter just to level set that Data Center expectation?

Jean Hu:

Yeah. So, I think for the second half, we said we expect Data Center business to grow approximately 50% versus first half. But right now, based on what we are seeing, we continue to see in that similar range of that 50%. So, we are very happy and pleased about the strong momentum of our Data Center business. On the GPU side, Lisa mentioned about $400 million, around $400 million […]”

The CEO let a comment slip that it would be “greater than $2 billion” for next year, and given how careful this management team is, I do think it’s important to note here:

Lisa Su:

Sure, Aaron. So, we've been planning the supply chain for the last year and we're always planning for success. So, certainly, for the current forecast of greater than $2 billion, we have adequate supply. But we have also planned for a supply chain forecast that could be significantly higher than that, and we would continue to work with customers to build that out.

For our purposes, as far as where this can go, the management team goes back to quoting 50% CAGR in this segment over the next few years: “So, I think we are big believers in the strength of the market. We previously said we believe that the compound annual growth rate could be 50% over the next three or four years.”

The March quarter is expected to be lumpy due to seasonality with the Client segment, Gaming, Embedded and then moving from El Capitan’s $400M in GPU revenue to $400M in Q1 from AI-driven GPU revenue (hyperscalers).

In regards to AMD’s AI-driven Client segment, here is a comment that helps solidify some of what our earlier analysis this year has discussed. From the CEO today: “What I'm most excited about in PCs is actually the AI PC. I think the AI PC opportunity is an opportunity to redefine what PCs are in terms of productivity tool and really sort of operating on sort of user data. And so, I think we're at the beginning of a wave there. We're investing heavily in Ryzen AI and the opportunity to really broaden sort of the AI capabilities of PCs going forward.”

Conclusion:

Gaming and Embedded will overshadow 4th Gen EPYC and also GPU sales in the near-term. However, we will be looking exclusively at the acceleration in the data center (EPYC and Instinct) and client segment (Ryzen AI) as the thesis. The market wanted to sell this report, yet AMD’s management team stood firm on their data center guide. For our purposes, some important hurdles were cleared – Q4 will come in as expected on the MI300s, Q1 will be the start of hyperscaler GPU sales, the margins will get better over time, and the Client segment is rebounding while also having its own AI story. Most importantly, we wanted to see AMD gain traction with hyperscalers, and that was confirmed this quarter for Q1.

Technically speaking, this report was stronger than Nvidia’s was this time last year despite the GPUs shipping at the same time seasonally – NVDA had 1% QoQ growth for data center revenue in the October quarter and then declined (-6%) QoQ in the Q4 January quarter. To jog your memory, it was the guide provided in May (for the July quarter) when Nvidia had its blowout quarter.

I don’t think AMD will take the exact path as Nvidia’s near-monopoly, rather, as was stated in the pre-earnings writeup, our goal is to see how this unfolds.

 Recommended Readings:

• Super Micro Fiscal Q1 Pre-Earnings: AI Marches Onward
• AMD Q3 Pre-Earnings: With Bated Breath for Q4 Commentary
• Meta Q3 Earnings Update
• Alphabet: Search Accelerates While Cloud Decelerates
• TSM Results: Recovery in sight but technicals look weak