Dell reported surging demand in AI optimized servers in Q1 with orders of $12.1 billion. This outpaces the entirety of last year while representing a 612% sequential increase from $1.7B last quarter. To further compare, the peak quarter for orders last year was $3.6B.
Server shipments were low in Q1, which has readthrough to Blackwell as all around Q1 was slow for Blackwell servers. Server shipments were guided to be nearly 4x higher sequentially in Q2. Check out the graph below for the sudden inflection point from this quarter in terms of server growth.
This strong AI server shipment forecast contributed to a nearly $4 billion beat for Q2’s guidance. Notably, Dell did not raise its revenue forecast for the year, suggesting that tariff-related impacts may still bite in H2, or that AI server shipments will be lumpy and not be linear from here out.
Cash flows improved significantly in Q1, with strong triple-digit YoY growth for both operating and free cash flow. Adjusted EPS missed consensus estimates by more than 8%. Despite misisng estimages, EPS increased 17% or 3X faster than revenue.
Revenue Growth to Accelerate 11 Points in Q2
Dell reported $23.38 billion in revenue in Q1, a slight <1% beat to estimates as all of its core businesses grew in the quarter. Revenue growth decelerated to 5.1% YoY in the quarter with Dell forecasting a sharp acceleration in Q2 as it is now rapidly ramping AI server shipments after orders surged in Q1.
For Q2, Dell guided $28.5 to $29.5 billion in revenue, or 15.9% YoY growth at the $29 billion midpoint, which marks a nearly 11-point sequential acceleration. Interestingly, while Q2’s guidance was nearly $4 billion ahead of the consensus estimate for 0.9% growth to $25.26 billion in revenue, Dell opted to maintain its FY26 revenue forecast at $101 to 105 billion.
Key Operating Segments
Infrastructure Solutions Group
Dell’s ISG segment grew 12% YoY to $10.32 billion in revenue, as Dell’s surge in AI server orders did not translate into booked revenue this quarter. ISG operating income was $0.99 billion, up 36% YoY for a 9.7% operating margin. This was up 1.7 points YoY in what is typically the lowest seasonal quarter for profitability for the segment.
In Q1, AI server orders were up 612% YoY and QoQ to $12.1 billion with Dell saying this exceeded the entirety of their fiscal 2025 AI server shipments of $9.8 billion. This surge in orders brought Dell’s AI server backlog up to $14.4 billion, up from $4.1 billion in Q4.
However, Q1’s AI server shipments were just $1.8 billion, up just 6% YoY and down more than (14%) QoQ. This likely boils down to the timing of Blackwell’s ramp, as Dell projected more than $7 billion in shipments in Q2.
This also raises the question that Q2’s shipment forecast is simply a surge aligning with Blackwell’s strong ramp, as Nvidia’s earnings pointed toward yesterday. With that said, AI server shipments could normalize at a much lower quarterly run rate, something in the range of $3.5 billion to $4 billion (with what we know today). Dell is remaining conservative by only slightly changing the language of its AI server guidance for the year, previously sticking to $15 billion but now aiming for $15B+.
Within ISG:
Servers and Networking revenue grew 16% YoY to $6.32 billion, with Dell stating that demand has grown for a sixth consecutive quarter, with traditional servers seeing double digit growth. Revenue has continued to decelerate off of Q2 FY25’s peak at 80%, though the $7 billion AI server shipment forecast will reverse this trend.
Storage revenue increased 6% YoY to $4.0 billion, the third consecutive quarter of storage growth.
Client Solutions Group
Client Solutions segment revenue fared much better than expected despite weak Consumer revenue, with growth of 5% YoY to $12.51 billion. This accelerated from 1% growth in Q4 and marked Dell’s second consecutive quarter for growth. CSG operating income was $653 million, down (16%) YoY for a 5.2% margin.
This growth was driven by increased momentum in Dell’s Commercial PC business, where Dell said that demand was up YoY for a fifth consecutive quarter and up double-digits this quarter. Management also said that they are “seeing clear indications that the install base is upgrading to new Windows 11 PCs, many of them AI PCs.”
Commercial’s momentum helped offset Consumer weakness, as revenue decelerated further, dropping (23%) QoQ and (19%) YoY, versus a (12%) YoY decline in Q4. Within CSG:
Commercial revenue accelerated from 5% in Q4 to 9% in Q1 to $11.05 billion.
Consumer revenue declined (19%) YoY to $1.46 billion.
Margins Down Sequentially on Seasonality, Mixed YoY
Dell had forecast its adjusted margins to come down sequentially due to seasonality, which played out in the quarter. On a YoY basis, gross margins contracted though margins down the line were relatively strong, suggesting Dell is beginning to capture some operating leverage via cost cuts.
GAAP gross margin was 21.1%, down half a point YoY and more than 2.5 points sequentially due to lower storage and higher AI server mix. Adjusted gross margin was 21.6%, down more than half a point YoY and more than 2.5 points sequentially.
GAAP operating margin was 5.0%, up more than 0.8 points YoY as Dell cut expenses by (3%) YoY, but down 4 points sequentially on seasonality. Adjusted operating margin was 7.1%, up half a point YoY but down 4 points sequentially.
GAAP net margin was 4.1%, lower than last year’s 4.5%. Adjusted net margin was 4.6%, up 0.3 points from 4.3% last year.
Adjusted EPS Misses Estimates, Though FY EPS Guide Raised
Dell reported a fairly large (8.3%) miss on adjusted EPS, reporting $1.55 in the quarter versus its guidance for $1.65 and analyst estimates for $1.69. However, Dell raised its FY26 EPS guidance, speaking to management’s confidence in executing as AI servers ramp and tariffs cloud the macro outlook.
For Q2, Dell guided for $2.15 to $2.35 in adjusted EPS for growth of 15% at midpoint, marking a slight deceleration from the 17.4% growth reported in Q1. Q3 and Q4 are expected to see EPS growth decelerate a bit further, with growth of just 10.7% in Q4.
For the full year, Dell slightly raised its FY26 adjusted EPS guidance to $9.40 for 15% growth, up from its prior view for $9.30 for 14% growth. Dell also slightly hiked its GAAP EPS view for FY26, now seeing $7.99 for 25% growth versus its prior view for $7.85 for 23% growth.
Cash Flows Show Strong Triple Digit Growth
Some of the stronger numbers of the report aside from AI server orders were Dell’s cash flow metrics, showing strong triple digit growth and a return to double digit margins for OCF.
Operating cash flow rose 168% YoY to $2.80 billion. OCF margin was 12.0%, up more than 7 points from 4.7% a year ago and more than 9.5 points higher than Q4’ s 2.4% margin.
Free cash flow rose 388% YoY to $2.23 billion, while adjusted free cash flow rose 258% YoY to $2.23 billion. FCF and adjusted FCF margin was 9.5%, a significant improvement from 2.1% and 2.8% a year ago.
Cash, equivalents and investments totaled $9.29 billion, up more than $4 billion QoQ. Debt also rose more than $4 billion QoQ to $28.78 billion.
Inventories were $7.42 billion, up more than 10% sequentially.
Share Buybacks and Dividends:
In the recent quarter, Dell returned $2.4 billion to shareholders with 22.1 million shares of stock repurchased and also paid a dividend of $0.53 per share. The CFO pointed out that since the start of 2023, Dell has returned $13.2 billion to shareholders through stock repurchases and dividends.
Earnings Q&A:
Lumpy Server Orders – Not Budging on the $15B Annual Forecast
Per our analysis yesterday, Nvidia stated the ramp for Blackwell is happening very quickly “On average, major hyperscalers are each deploying nearly 1,000 NVL72 racks or 72,000 Blackwell GPUs per week and are on track to further ramp output this quarter.”
Dell primarily focuses on Tier 2 CSPs, yet a press release was issued stating the following that would indicate Dell is shipping Nvidia’s largest systems at scale: “One of Dell’s U.S. factories can ship thousands of NVIDIA Blackwell GPUs to customers in a week. It’s why they were chosen by one of their largest customers to deploy 100,000 NVIDIA GPUs in just six weeks.”
Also buried in the call was a comment by the CEO stating they have 3,000 AI customers – which feels very high to me given the current order number (meaning orders should follow i time): “Our enterprise growth is exciting with over 3,000 customers now buying various forms of our Dell AI factories. We saw a mix from Hopper technology and Blackwell technology across those. We saw it with [Worm and x86] (ph), so a great cross-representation there.”
Despite this excitement, Dell offered a muted tone on the call especially as they declined to increase their AI forecast for $15 billion in AI servers this year, stating: “The customer deployments that we have in front of us are large. They're complex. They have very detailed schedule deliveries. There's lots of dependencies on this. We've talked about this business being lumpy and nonlinear. The dependencies in this business are waiting for data centers to be built, power to be provided, direct liquid cooling infrastructure put in place. We're orchestrating a highly complex supply chain [..]” and later it was stated: “[…] I like our prospects of converting more pipeline in the second half, but at this point, we're on the $15 billion plus side. Our annual guidance that we just delivered suggests that's exactly where we are.”
My readthrough is that Dell is not willing to offer guidance on these systems that have had many delays. From what I can tell, companies in the United States supply chain would rather just surprise the market down the line than overpromise on something outside of their control.
When it comes to other Nvidia server makers such as Foxconn, Wistron and Quanta, many of them have extensive manufacturing operations in China (although headquartered in Taiwan). I would not be surprised if we see Nvidia more “encouraged” to use USA-based server makers such as Dell somewhere down the line.
Discussions around AI Server Margins
There was an exchange around ISG margins on the call with an analyst trying to pinpoint if AI servers result in “a low single-digit operating margin” — however, management pushed back on this stating that ISG margins are expanding on a QoQ basis due to AI servers.
Because AI server margins can make or break a stock like Dell (or Supermicro), I’m quoting the response in full – this is about the guide and not the current quarter results, which had lower margins.
The CEO stated: “When I think about AI and the numbers that we gave, the $7 billion of incremental revenue. When you look at it, I believe it's roughly $4 billion on a year-over-year basis. It's roughly $5 billion on a quarter-over-quarter basis. And it drives significant gross margin dollar growth on both a year-over-year and quarter-over-quarter basis. And it drives significant operating income dollar growth on a quarter-over-quarter and year-over-year basis. I'll turn it over to Yvonne, she can add more.”
The CFO later stated: “Yes, I'd say embedded within the guide is a 10% quarter-over-quarter increase in gross margin dollars. As Jeff mentioned, we're seeing — what we're seeing in ISG quarter-over-quarter is [indiscernible] $5.3 billion more revenue, with roughly $0.5 billion more in operating income, which is being driven by AI server profitability and to a lesser extent improvement in the profitability within our Storage portfolio.”
This would mean the flat guide on adjusted operating margin is coming from traditional servers as it was stated in the opening remarks: “We are expecting sub-seasonal performance in traditional server and storage, our larger profit pools that provide scale, as customers evaluate their IT spend for the year given the dynamic macro environment.”
Weak Macro Backdrop:
When asked directly about a tariff pull forward, Dell cut to the chase and confirmed they believe a pull forward occurred in all three of their traditional businesses of PCs, storage and non-AI servers.
The CEO was quite granular in discussing the details:
“Jeff Clarke:
I think I mentioned in my remarks at North American, EMEA and APJ all grew double digits from a demand perspective. We did see a slowdown in month three. Month one was greater than January. Month two was greater than February. Month three slowed in weeks 10 through 12 in actually all three US businesses, commercial PCs, traditional servers, and storage. So clearly there is a bump along the journey there, along they reference to that with a slowdown in our traditional server business. So, a pull ahead. We are still optimistic about the year. We have all of the businesses growing. We are maybe a little more needed in what we think in those businesses. They may be down a point in terms of their absolute market growth. I don't think anyone knows, but when you look at traditional servers, you look at storage, the other two businesses that I was referring to. I think, both were growing nicely. North America speed bump with 10 through 12 in month three. Worked our way through that. We now have that reflected in our guidance in Q2.”
Conclusion:
We took a stab at Dell as it was becoming apparent from Nvidia’s report that Dell would likely beat its AI server business. It was a blowout in that regard, as the chart shows, yet there are a few puts and takes to consider. First off, Dell is not comfortable guiding to more than what they are sure will be delivered in light of many delays with Blackwell. Secondly, Dell’s other segments saw a pull forward and will weigh on results as we move into future quarters. However, one reason we took a stab is that a bullish setup is forming – like with many AI stocks right now, it requires some patience.
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.
Nvidia posted a strong Q1 and marginally missed estimates in Q2 due to Blackwell revenue that exceeded expectations. The larger Blackwell systems are in full production and are shipping in volume now, which sets up a strong second half of the year.
According to management commentary, the ramp is happening very quickly: “On average, major hyperscalers are each deploying nearly 1,000 NVL72 racks or 72,000 Blackwell GPUs per week and are on track to further ramp output this quarter.” The rough math here implies hyperscalers are deploying $3 billion every week right now since each rack goes for $3 million. Furthermore, the run rate of this comment implies data center revenue will be above and beyond analyst consensus for Q2, Q3 and Q4 – thus, either analyst consensus comes up or these systems will become further supply constrained somewhere down the line and analysts are being conservative for now.
Among the many reasons that Blackwell is an improvement compared to the Hopper architecture, management focused on inference stating: “Compared to Hopper, Grace Blackwell is some 40 times higher speed and throughput compared.”
The loss of China revenue in Q1 was $2.5 billion yet inventory charges were higher at $4.5 billion from orders placed prior to April 9. For Q2, the loss of China revenue was $8 billion –or about $2 to $3B higher than the typical $5.5B in China revenue. The readthrough is that Blackwell came in $2-$3B above analyst expectations to absorb that impact from China, since guidance marginally missed.
You can view an interview on Fox where I discussed the puts and takes going into the earnings report plus the new price target I/O Fund published here. If you’re brief on time, the takeaway is that Blackwell has enough ammo to push the stock into the mid-to-high $200s or a $6+ trillion market cap. I discuss this and more below.
Slight Revenue Beat in Q1, Marginal Miss in Q2
Nvidia reported a slight revenue beat in Q1, reporting 69.2% YoY growth to $44.06 billion in revenue, just ahead of the $43.25 billion consensus.
For Q2, Nvidia guided $45 billion, +/- 2%, representing a deceleration to 49.8% YoY growth and a marginal miss at midpoint versus consensus at $45.66 billion. At the low end of the guide, revenue growth would be up only 2.3% sequentially, reflecting how large of an impact the H20 ban is having on growth.
Nvidia also quantified more of the H20 impact, providing details on the revenue impact to both Q1 and Q2 – in total, both quarters are seeing a combined impact of just over $15 billion. Nvidia added that the inventory charge of $4.54 billion was less than the $5.5 billion anticipated as it was able to re-use certain materials.
For Q1, Nvidia said that it recorded $4.6 billion in H20 revenue, or about 10.4% of revenue, while it was unable to ship an additional $2.5 billion of H20 revenue due to export restriction. In total, this implies $7.1 billion in H20 revenue in Q1. This would represent around 16% of total revenue or 18.2% of data center revenue in the quarter.
For Q2, Nvidia said that its guidance reflects the loss of approximately $8 billion in H20 sales, implying nearly 13% QoQ growth was expected to fill extremely high Chinese demand for the chip. However, based off management’s commentary, its $45 billion guide for Q2 suggest that other Blackwell SKUs are ramping rapidly and filling much of the H20 void.
Key Segments
Data Center
Nvidia reported 73.3% growth in data center revenue to $39.11 billion in Q1, marginally higher than analyst expectations from Visible Alpha of $39.08 billion. This marked the end of Nvidia’s seven-quarter streak of $1 billion-plus beats in the segment – based on the Visible Alpha estimate, Nvidia beat by just $33 million, its lowest in the past nine quarters.
Compute revenue rose 76% YoY but just 5% QoQ to $34.16 billion, impacted by the H20 ban, while Networking revenue rebounded swiftly, rising 56% YoY and 65% QoQ to $4.96 billion. Nvidia said Networking’s performance was “driven by the growth of NVLink compute fabric in our GB200 systems and continued adoption of Ethernet for AI solutions at cloud service providers and consumer internet companies.”
In contrast to the prior two quarters, Nvidia did not give a number for Blackwell revenue in the quarter, stating only that its Blackwell ramp expanded to all customer categories and that large CSPs remained its largest customers at just under 50% of data center revenue.
Nvidia also said that its hyperscaler customers “are each deploying nearly 1,000 NVL 72 racks or 72,000 Blackwell GPUs per week,” with this output level on track to ramp further this quarter.
Gaming revenue rebounded sharply, rising 48% QoQ and accelerating 53 points sequentially to 42% YoY with revenue of $3.76 billion in Q1. Nvidia said that this was driven by its Blackwell architecture and the fastest ramp in company history.
Automotive revenue rose 72% YoY but declined (1%) QoQ to $567 million.
Pro Viz revenue rose 19% YoY and was approximately flat QoQ at $509 million.
OEM and Other revenue rose 42% YoY but declined (12%) QoQ to $111 million.
Margins Take Large Hit from H20, Though Q2 Points to Swift Rebound
Nvidia’s margins took a rather large hit from the H20-related inventory write-down, with gross margin and operating margins contracting significantly. However, management’s guidance for Q2 points to a rapid recovery in margins as Blackwell ramps, likely aided by its pricing power.
GAAP gross margin was 60.5% and adjusted gross margin was 61%, around 10 points below management’s initial guidance for 70.6% and 71% due to the $4.54 billion charge related to the H20 ban. Management noted that excluding the charges associated with the ban, adjusted gross margin would’ve been 71.3%, at the upper end of the guided range of 71% +/- 0.5%.
For Q2, management guided for 71.8% GAAP gross margins and 72% adjusted gross margins, a rebound of approx. 11 points sequentially.
GAAP operating margin was 49.1%, well below guidance for 58.5% and a sequential contraction of 12 points. Adjusted gross margin was 52.8%, nearly 10 points below the guide for 62.6% and a sequential contraction of more than 12 points.
For Q2, management’s guidance implies operating margins will rebound with gross margins, projecting approximately a 10 point sequential expansion to a 59.1% GAAP and 63.1% adjusted operating margin.
GAAP net margin was 42.6%, while adjusted net margin was 45.2%. The broad-based margin recovery in Q2 is expected to mostly transfer through to the bottom line, with management guiding for a 7.6 point recovery to a 50.2% GAAP net margin.
EPS Beats, Growth Expected to Rebound
Nvidia reported a slight EPS beat despite the margin contractions, with adjusted EPS of $0.81 coming in ahead of the $0.75 estimate. GAAP EPS of $0.76 missed estimates for $0.81.
Adjusted EPS growth slowed quite dramatically, decelerating more than 38 points sequentially, in part due to the H20 ban; Nvidia noted that excluding the ban, adjusted EPS would be $0.96. This would represent YoY growth of 57.4% versus the 32.8% reported.
Looking ahead, adjusted EPS growth is expected to rebound and remain in the low to mid-40% range as margins recover. However, given that Q1’s EPS excluding the ban showed growth in the high-50% range, estimates may move higher as Q2’s margin outlook shows almost no persisting impact.
Cash Flows and Balance Sheet
Cash flows were surprisingly strong as Nvidia’s cash flow margins expanded approximately 20 points sequentially, while it added more than $10 billion in cash to its balance sheet.
Operating cash flow was $27.41 billion, up nearly 79% YoY on higher revenue, timing of its cash collections, and lower cash taxes. OCF margin was 62.2%, up 20 points QoQ and more than 3 points YoY. Nvidia said it expects a substantial increase in cash taxes in Q2, which will weigh on OCF.
Free cash flow was $26.14 billion, up 75% YoY. FCF margin was 59.3%, up nearly 20 points QoQ and just 2 points YoY.
Inventories were $11.33 billion, rising more than 12% QoQ. However, days sales of inventory decreased from 86 days in Q4 to 59 days in Q1, due to the sharp increase in COGS from the H20 inventory charges.
Accounts receivable were $22.1 billion, declining just over (4%) QoQ. Days sales outstanding decreased sequentially from 53 days to 46 days due to improved shipment linearity (shipments more evenly distributed throughout the quarter) and timing of collections.
Cash and equivalents rose more than $10 billion sequentially to $53.69 billion, despite Nvidia returning more than $14.3 billion to shareholders in the quarter with $14.1 billion in share repurchases.
Debt remained steady at $8.46 billion.
Earnings Q&A:
Inference Demand is Skyrocketing
In the opening remarks, management stated they are seeing “a sharp jump in inference demand.” Our firm recently covered the 5X increase in tokens quoted by Microsoft to 50T tokens per month and 100T per quarter stated in their most recent earnings report. In that analysis, we pointed toward up to $18 billion in annualized revenue for API usage in high-end models. Google recently stated at their I/O Developer event they are processing 450T tokens per month up 50X from a year ago.
In the opening remarks the following was shared about the NVL72 inferencing capabilities: “Inference serving startups are now serving models using B200, tripling their token generation rate and corresponding revenues for high-value reasoning models such as DeepSeek-R1 as reported by artificial analysis. NVIDIA Dynamo on Blackwell NVL72 turbocharges AI inference throughput by 30x for the new reasoning models, sweeping the industry […] In the latest MLPerf Inference results, we submitted our first results using GB200 NVL72, delivering up to 30x higher inference throughput compared to our 8-GPU H200 submission on the challenging Llama 3.1 benchmark.”
Later, in the Q&A session, Jensen Huang stated inference is reaching an inflection point, stating “we've reached an extraordinary milestone with AIs that are reasoning, are thinking, what people call inference time scaling. Of course, it created a whole new — we've entered an era where inference is going to be a significant part of the compute workload.”
It was then re-emphasized again in the Q&A with Huang stating:
“Yeah, thanks. Thanks, Ben. I would say compared to the beginning of the year, compared to GTC timeframe, there are four positive surprises. The first positive surprise is the step function demand increase of reasoning AI, I think it is fairly clear now that AI is going through an exponential growth, and reasoning AI really busted through [… So, number one is inference reasoning and the exponential growth there, demand growth.”
Note on Valuation:
I've written a substantial amount on Blackwell — perhaps the most important being my $10T prediction came out about two months after Blackwell was announced at GTC. If you read-between-the-lines on our new price target, then I’m saying Nvidia can reach more thn a $6T market cap as soon as next year.
This relies on two assumptions. The first is that we see Nvidia reach the valuation it saw in 2024 when the forward PE Ratio was at 50 forward a handful of times. That implies a move of up to 51% as it stands today.
On the top line, NVDA has traded as high as 28 forward PS, implying room of up to 75%.
The second assumption is that Nvidia beats estimates, forcing the valuation higher. We’ve seen a small glimpse of this in Q2 with $2 to $3B in Blackwell revenue absorbing China losses. However, I’ve been crystal clear that it’s the August call and November call that will be fireworks. I expect to see Nvidia grand slam beats in these quarters especially and/or analyst consensus moving up into those quarters, creating more room in the valuation then the 50% to 75% we see currently.
Conclusion:
The marginal miss in Q2 would have been more pronounced if Blackwell were not ramping. Pay attention, as this is the cyclical bottom for Nvidia as the Hopper generation fades out and yet its earnings reports have been unscathed. The future is bright and the fireworks are locked and loaded for H2.
Please note: The I/O Fund conducts research and draws conclusions for the company’s portfolio. We then share that information with our readers and offer real-time trade notifications. This is not a guarantee of a stock’s performance and it is not financial advice. Please consult your personal financial advisor before buying any stock in the companies mentioned in this analysis. Beth Kindig and the I/O Fund own shares in NVDA at the time of writing and may own stocks pictured in the charts.
Nvidia posted a strong Q1 and marginally missed estimates in Q2 due to Blackwell revenue that exceeded expectations. The larger Blackwell systems are in full production and are shipping in volume now, which sets up a strong second half of the year.
According to management commentary, the ramp is happening very quickly: “On average, major hyperscalers are each deploying nearly 1,000 NVL72 racks or 72,000 Blackwell GPUs per week and are on track to further ramp output this quarter.” The rough math here implies hyperscalers are deploying $3 billion every week right now since each rack goes for $3 million. Furthermore, the run rate of this comment implies data center revenue will be above and beyond analyst consensus for Q2, Q3 and Q4 – thus, either analyst consensus comes up or these systems will become further supply constrained somewhere down the line and analysts are being conservative for now.
Among the many reasons that Blackwell is an improvement compared to the Hopper architecture, management focused on inference stating: “Compared to Hopper, Grace Blackwell is some 40 times higher speed and throughput compared.”
The loss of China revenue in Q1 was $2.5 billion yet inventory charges were higher at $4.5 billion from orders placed prior to April 9. For Q2, the loss of China revenue was $8 billion –or about $2 to $3B higher than the typical $5.5B in China revenue. The readthrough is that Blackwell came in $2-$3B above analyst expectations to absorb that impact from China, since guidance marginally missed.
You can view an interview on Fox where I discussed the puts and takes going into the earnings report plus the new price target I/O Fund published here. If you’re brief on time, the takeaway is that Blackwell has enough ammo to push the stock into the mid-to-high $200s or a $6+ trillion market cap. I discuss this and more below.
Slight Revenue Beat in Q1, Marginal Miss in Q2
Nvidia reported a slight revenue beat in Q1, reporting 69.2% YoY growth to $44.06 billion in revenue, just ahead of the $43.25 billion consensus.
For Q2, Nvidia guided $45 billion, +/- 2%, representing a deceleration to 49.8% YoY growth and a marginal miss at midpoint versus consensus at $45.66 billion. At the low end of the guide, revenue growth would be up only 2.3% sequentially, reflecting how large of an impact the H20 ban is having on growth.
Nvidia also quantified more of the H20 impact, providing details on the revenue impact to both Q1 and Q2 – in total, both quarters are seeing a combined impact of just over $15 billion. Nvidia added that the inventory charge of $4.54 billion was less than the $5.5 billion anticipated as it was able to re-use certain materials.
For Q1, Nvidia said that it recorded $4.6 billion in H20 revenue, or about 10.4% of revenue, while it was unable to ship an additional $2.5 billion of H20 revenue due to export restriction. In total, this implies $7.1 billion in H20 revenue in Q1. This would represent around 16% of total revenue or 18.2% of data center revenue in the quarter.
For Q2, Nvidia said that its guidance reflects the loss of approximately $8 billion in H20 sales, implying nearly 13% QoQ growth was expected to fill extremely high Chinese demand for the chip. However, based off management’s commentary, its $45 billion guide for Q2 suggest that other Blackwell SKUs are ramping rapidly and filling much of the H20 void.
Key Segments
Data Center
Nvidia reported 73.3% growth in data center revenue to $39.11 billion in Q1, marginally higher than analyst expectations from Visible Alpha of $39.08 billion. This marked the end of Nvidia’s seven-quarter streak of $1 billion-plus beats in the segment – based on the Visible Alpha estimate, Nvidia beat by just $33 million, its lowest in the past nine quarters.
Compute revenue rose 76% YoY but just 5% QoQ to $34.16 billion, impacted by the H20 ban, while Networking revenue rebounded swiftly, rising 56% YoY and 65% QoQ to $4.96 billion. Nvidia said Networking’s performance was “driven by the growth of NVLink compute fabric in our GB200 systems and continued adoption of Ethernet for AI solutions at cloud service providers and consumer internet companies.”
In contrast to the prior two quarters, Nvidia did not give a number for Blackwell revenue in the quarter, stating only that its Blackwell ramp expanded to all customer categories and that large CSPs remained its largest customers at just under 50% of data center revenue.
Nvidia also said that its hyperscaler customers “are each deploying nearly 1,000 NVL 72 racks or 72,000 Blackwell GPUs per week,” with this output level on track to ramp further this quarter.
Gaming revenue rebounded sharply, rising 48% QoQ and accelerating 53 points sequentially to 42% YoY with revenue of $3.76 billion in Q1. Nvidia said that this was driven by its Blackwell architecture and the fastest ramp in company history.
Automotive revenue rose 72% YoY but declined (1%) QoQ to $567 million.
Pro Viz revenue rose 19% YoY and was approximately flat QoQ at $509 million.
OEM and Other revenue rose 42% YoY but declined (12%) QoQ to $111 million.
Margins Take Large Hit from H20, Though Q2 Points to Swift Rebound
Nvidia’s margins took a rather large hit from the H20-related inventory write-down, with gross margin and operating margins contracting significantly. However, management’s guidance for Q2 points to a rapid recovery in margins as Blackwell ramps, likely aided by its pricing power.
GAAP gross margin was 60.5% and adjusted gross margin was 61%, around 10 points below management’s initial guidance for 70.6% and 71% due to the $4.54 billion charge related to the H20 ban. Management noted that excluding the charges associated with the ban, adjusted gross margin would’ve been 71.3%, at the upper end of the guided range of 71% +/- 0.5%.
For Q2, management guided for 71.8% GAAP gross margins and 72% adjusted gross margins, a rebound of approx. 11 points sequentially.
GAAP operating margin was 49.1%, well below guidance for 58.5% and a sequential contraction of 12 points. Adjusted gross margin was 52.8%, nearly 10 points below the guide for 62.6% and a sequential contraction of more than 12 points.
For Q2, management’s guidance implies operating margins will rebound with gross margins, projecting approximately a 10 point sequential expansion to a 59.1% GAAP and 63.1% adjusted operating margin.
GAAP net margin was 42.6%, while adjusted net margin was 45.2%. The broad-based margin recovery in Q2 is expected to mostly transfer through to the bottom line, with management guiding for a 7.6 point recovery to a 50.2% GAAP net margin.
EPS Beats, Growth Expected to Rebound
Nvidia reported a slight EPS beat despite the margin contractions, with adjusted EPS of $0.81 coming in ahead of the $0.75 estimate. GAAP EPS of $0.76 missed estimates for $0.81.
Adjusted EPS growth slowed quite dramatically, decelerating more than 38 points sequentially, in part due to the H20 ban; Nvidia noted that excluding the ban, adjusted EPS would be $0.96. This would represent YoY growth of 57.4% versus the 32.8% reported.
Looking ahead, adjusted EPS growth is expected to rebound and remain in the low to mid-40% range as margins recover. However, given that Q1’s EPS excluding the ban showed growth in the high-50% range, estimates may move higher as Q2’s margin outlook shows almost no persisting impact.
Cash Flows and Balance Sheet
Cash flows were surprisingly strong as Nvidia’s cash flow margins expanded approximately 20 points sequentially, while it added more than $10 billion in cash to its balance sheet.
Operating cash flow was $27.41 billion, up nearly 79% YoY on higher revenue, timing of its cash collections, and lower cash taxes. OCF margin was 62.2%, up 20 points QoQ and more than 3 points YoY. Nvidia said it expects a substantial increase in cash taxes in Q2, which will weigh on OCF.
Free cash flow was $26.14 billion, up 75% YoY. FCF margin was 59.3%, up nearly 20 points QoQ and just 2 points YoY.
Inventories were $11.33 billion, rising more than 12% QoQ. However, days sales of inventory decreased from 86 days in Q4 to 59 days in Q1, due to the sharp increase in COGS from the H20 inventory charges.
Accounts receivable were $22.1 billion, declining just over (4%) QoQ. Days sales outstanding decreased sequentially from 53 days to 46 days due to improved shipment linearity (shipments more evenly distributed throughout the quarter) and timing of collections.
Cash and equivalents rose more than $10 billion sequentially to $53.69 billion, despite Nvidia returning more than $14.3 billion to shareholders in the quarter with $14.1 billion in share repurchases.
Debt remained steady at $8.46 billion.
Earnings Q&A:
Inference Demand is Skyrocketing
In the opening remarks, management stated they are seeing “a sharp jump in inference demand.” Our firm recently covered the 5X increase in tokens quoted by Microsoft to 50T tokens per month and 100T per quarter stated in their most recent earnings report. In that analysis, we pointed toward up to $18 billion in annualized revenue for API usage in high-end models. Google recently stated at their I/O Developer event they are processing 450T tokens per month up 50X from a year ago.
In the opening remarks the following was shared about the NVL72 inferencing capabilities: “Inference serving startups are now serving models using B200, tripling their token generation rate and corresponding revenues for high-value reasoning models such as DeepSeek-R1 as reported by artificial analysis. NVIDIA Dynamo on Blackwell NVL72 turbocharges AI inference throughput by 30x for the new reasoning models, sweeping the industry […] In the latest MLPerf Inference results, we submitted our first results using GB200 NVL72, delivering up to 30x higher inference throughput compared to our 8-GPU H200 submission on the challenging Llama 3.1 benchmark.”
Later, in the Q&A session, Jensen Huang stated inference is reaching an inflection point, stating “we've reached an extraordinary milestone with AIs that are reasoning, are thinking, what people call inference time scaling. Of course, it created a whole new — we've entered an era where inference is going to be a significant part of the compute workload.”
It was then re-emphasized again in the Q&A with Huang stating:
“Yeah, thanks. Thanks, Ben. I would say compared to the beginning of the year, compared to GTC timeframe, there are four positive surprises. The first positive surprise is the step function demand increase of reasoning AI, I think it is fairly clear now that AI is going through an exponential growth, and reasoning AI really busted through [… So, number one is inference reasoning and the exponential growth there, demand growth.”
Note on Valuation:
I've written a substantial amount on Blackwell — perhaps the most important being my $10T prediction came out about two months after Blackwell was announced at GTC. If you read-between-the-lines on our new price target, then I’m saying Nvidia can reach more thn a $6T market cap as soon as next year.
This relies on two assumptions. The first is that we see Nvidia reach the valuation it saw in 2024 when the forward PE Ratio was at 50 forward a handful of times. That implies a move of up to 51% as it stands today.
On the top line, NVDA has traded as high as 28 forward PS, implying room of up to 75%.
The second assumption is that Nvidia beats estimates, forcing the valuation higher. We’ve seen a small glimpse of this in Q2 with $2 to $3B in Blackwell revenue absorbing China losses. However, I’ve been crystal clear that it’s the August call and November call that will be fireworks. I expect to see Nvidia grand slam beats in these quarters especially and/or analyst consensus moving up into those quarters, creating more room in the valuation then the 50% to 75% we see currently.
Conclusion:
The marginal miss in Q2 would have been more pronounced if Blackwell were not ramping. Pay attention, as this is the cyclical bottom for Nvidia as the Hopper generation fades out and yet its earnings reports have been unscathed. The future is bright and the fireworks are locked and loaded for H2.
Please note: The I/O Fund conducts research and draws conclusions for the company’s portfolio. We then share that information with our readers and offer real-time trade notifications. This is not a guarantee of a stock’s performance and it is not financial advice. Please consult your personal financial advisor before buying any stock in the companies mentioned in this analysis. Beth Kindig and the I/O Fund own shares in NVDA at the time of writing and may own stocks pictured in the charts.
It is easy to draw on one’s emotional bias and therefore build a believable case for what the market will do next. We think this is a mistake for investors positioning for the remainder of 2025. Instead, we will continue to let the markets tell us what is to come.
“The next move will be a corrective rally that makes a lower high. The targets for this bounce are between 5600 – 6050.”
We further stated that once we see our first larger correction from this region, how the market corrects from there will likely determine the remainder of the year.
Having an unbiased game plan still applies and continues to act as a balance beam in an emotionally charged market. We are seeing mounting evidence that this bounce may be the start of a new push to all-time highs, such as improved breadth, better than expected earnings plus the size of this bounce. However, one can’t ignore the unprecedented levels of uncertainty shown in key indexes, coupled with a growing problem in the U.S. bond market.
In this report, we’ll lay out the unbiased case for each scenario for our 2025 stock market outlook. We do this so that we can be aligned with the developing trend once it is revealed.
Why the S&P 500 Could Reach New All-Time Highs This Year
Fibonacci Retracements
There are several factors that suggest the current rally is not a bear market bounce. One of the most interesting facts is that we’ve never seen a bear market bounce retrace this much (and this quickly) without turning into a new uptrend.
The simplest method for measuring a bounce is to use Fibonacci Retracement levels. For those not familiar with this simple technique, when a market starts to bounce after a period of volatility, you simply divide the drop into key Fibonacci numbers. So, 38.2%, 50%, 61.8%, and 76.4% retracements of the drop are areas of interest.
Using this technique, we can get an idea of the size of the current bounce relative to what history says about bear market bounces. Going back to 1929 there have been 19 bear markets, as defined by a decline of 20% or more in the markets. Once a market dropped into bear market territory, we would usually see a bounce back to the 50% retracement level before starting to trend lower.
For example, the 2000 peak entered an official bear market by declining 20% in February of 2001. In late March, the market staged a 22% rally that just barely made it over the 50% retracement of the entire drop. It then turned lower and resumed the downtrend.
March 2001 saw a bear market rally of 22% that briefly surpassed the 50% retracement level before turning lower. Source: I/O Fund
The above scenario has been the most likely outcome for prior bear market bounces. However, of the 19 bear markets since 1929, there have only been four bear market bounces that made it to the 61.8% retracement – 1938, 1947, 2008, and the most recent bear market in 2022, which is shown below.
August 2022’s bear market rally was one of the rare bear market bounces that touched the 61.8% retracement level and turned lower. Source: I/O Fund
There is no instance, so far, where a bear market rally moved beyond the 61.8% retracement of the entire drop and was not the start of a new uptrend. The current bounce not only exceeded the 61.8% retracement level, but it also went above the 76.4% retracement level.
The rally so far in April and May 2025 has surpassed the 61.8% and 76.4% retracement levels. If it is a bear market rally, it will be the 1st ever to bounce this high. Source: I/O Fund
This shifts the probability that we are in a bear market bounce to being low, based on historic standards. It would not only be the first bear market bounce that exceeded the 61.8% retracement level, but it would be the first bear market to exceed the 76.4% retracement level.
Advance/Decline Line
The Advance-Decline Line (A/D Line) is a widely used indicator that tracks market breadth by showing how many stocks are rising versus falling on a given trading day. How it works is that each trading day, analysts tally the number of stocks that closed higher than the previous day (advancing) and subtract the number of stocks that closed lower (decline). This value is then added to the prior day’s cumulative A/D Line, creating a running total that reveals whether participation in the market is expanding or narrowing over time.
What is particularly interesting about the A/D line is how it tends to lead price coming out of periods of volatility. The chart below shows this phenomenon leading equities to new highs in 2016, 2018, and 2022.
Note how both the S&P 500 and the A/D line topped around the same periods in all four tops. However, the A/D Line has a history of breaking out to new highs months before the S&P 500 – in the case of the 2023 recovery, the Advance Decline line broke to new highs almost a year before the S&P 500.
The Advance Decline Line tends to lead the S&P 500 coming out of periods of volatility. It has been a reliable signal that new highs will follow. Source: I/O Fund
The reason this is important is because every instance the market has had a meaningful correction since the 2008 top, the Advance Decline line would breakout to new highs months before price, signaling that a new high is the broad market is likely to follow.
Today, we are seeing the same phenomenon. The A/D line broke to new highs on April 29th, while the S&P 500 remains below its February 19th high.
The Advance Decline line topped alongside the S&P 500 in February of 2025 but has since broke out to new highs. Will the S&P 500 follow? Source: I/O Fund
If history is a guide, seeing breadth, as measured by the Advance Decline line, break to new highs, suggests price will follow.
Earnings Growth Much Better than Expected in Q1
We usually do not see large and prolonged declines while earnings are growing. We tend to see a consistent pattern of misses in earnings that is accompanied with a clear deceleration. Based on current reports, earnings are coming in better than expected, with earnings growth for the S&P 500 rising as more companies report. The index is also on track to report its second consecutive quarter of double-digit earnings growth and seventh consecutive quarter of growth.
Data from LSEG I/B/E/S as of May 16 placed the S&P 500’s Q1 2025 blended EPS growth rate at 14.3% YoY with 92% of companies reporting. Earnings growth is up more than 4 points since April 25’s 10.1% blended growth rate and up more than 6 points since April 1’s 8.0% blended growth rate.
Q1’s blended earnings growth for the S&P 500 is expected to be 14.3%, up more than 6 points since April 1. Source: I/O Fund, data from LSEG I/B/E/S
Blended earnings growth estimates for the remainder of 2025 have come down rather sharply as the market digested April’s tariff announcement, with growth now expected to be in the mid-single digit range down from the strong double-digit range.
2026 earnings growth is estimated to be rather robust, accelerating to nearly 16% YoY by Q2 before moderating to the 14% range by Q4, per LSEG I/B/E/S data. Though we are not seeing a pattern of earnings misses this quarter, these growth rates could change quickly, as Q1 26’s growth estimate has already come down nearly 8 points in six weeks. There has also been a considerable number of discussions around tariff pull forwards, to where indecisive buyers rush to make purchases before tariffs take effect.
Risks That Cannot Be Ignored: The 30-Year Stock-Bond Correlation is Breaking Amid Record Market Uncertainty
Even though we are seeing some signals that historically precede higher stock prices, one can’t underestimate the backdrop of the unique risks associated with the current stock market. For one, markets do not like uncertainty. When uncertainty is introduced into equity valuations, we tend to see aggressive repricing of perceived risk within the markets. In other words, sell first and ask questions later. This is what happened during COVID, as well as Liberation Day.
Though fear has subsided due to the size of this bounce in the markets, it’s worth noting that we are seeing a record high in the indexes that measure geo-political and economic uncertainty. The Economic Policy Uncertainty Index (EPU), which provides a quantifiable measurement of global uncertainty based on news headlines, global conflicts, tariffs, and changing tax codes, is signaling the highest level of uncertainty seen in more than two decades.
The Economic Policy Uncertainty Index (EPU) is showing the highest level of uncertainty in over a century. Source: Economic Policy UncertaintyEconomic Policy Uncertainty
This is further backed up by the Bloomberg Trade Policy Index, which is also at record levels of uncertainty.
Trade policy uncertainty shot up to a record high in 2025. Source: Bloomberg EconomicsBloomberg Economics
To make matters more unsettling, when the markets enter a period of uncertainty, which increases market volatility, we tend to see a flight into long-duration government bonds – the tried-and-true haven. For over 30 years, when stocks go down, bonds go up, and this has been the pattern investors can count on, making a diversified portfolio of stocks and bonds the ideal instrument for weathering periods of volatility with ease.
Considering that we are seeing historic levels of uncertainty, coupled with heightened volatility, this correlation states that we should have seen a notable increase in government bonds, as investors turn toward safety. However, since the market peaked on February 19th, the ETF that tracks long dated government bonds, TLT, is down nearly 7%.
Some might suggest that the market is forward looking, and that bonds did not go higher because the market may be pricing in a full recovery. Once again, no one knows for sure, but if this is the case, then the same logic should also apply to prior periods of quick volatility – like 2010, 2011, 2015, and 2020. These were periods of uncertainty and heightened volatility that were short lived, yet while uncertainty was high during these periods, we saw investors flee into bonds, quickly pushing TLT up 25% to 53%, as shown in the chart below.
Bonds historically have moved inversely to the stock market during periods of uncertainty, though 2022 and 2025’s market saw bonds falling while stocks fall. Source: I/O Fund
Now, compare this to today’s market. We saw the S&P 500 drop into bear market territory in just over one month, with some of the highest recorded geo-political uncertainty on record. Fear and uncertainty were so elevated during this time that we saw the volatility index (VIX) post a closing price of 45. Since 1990, there have been only three periods where we saw the VIX close over this level – 2008, 2009, and 2020.
This suggests that we are potentially seeing a 30-year correlation between stocks and bonds shift in real-time. And, if this correlation-break persists, it will pose a much bigger risk to financial markets than tariffs or political uncertainty.
Dollar Weakness and Debt Maturity Crisis Could Force U.S. Rates Even Higher
Bonds appear to be setting up for a breakdown, not a breakout. In other words, investors should expect rates to go higher while the U.S. has to refinance $7 trillion (due now) of its $9.2 trillion in maturing debt this year, with another $5 trillion due next year. The $9.2 trillion alone from 2025 is around one-third of the market value of marketable Treasury debt, and nearly 30% of US GDP.
While higher rates loom over the economy and threaten to weigh on growth, as the 10-year and 30-year rise past 4.5% and 5%, there’s also broader implications to consumers and government spending. Higher rates will put upward pressure on borrowing costs, making mortgages, car loans, or variable-rate-based loans including credit cards more expensive.
Net interest payments on debt are surging, with 2025’s estimated payments at $952 billion, up 8% YoY, and more than 175% higher since 2020. Interest payments are expected to surpass $1 trillion as soon as 2026. From 2025 to 2035, net interest payments are currently forecast to total $13.8 trillion cumulatively.
The US’ net interest payments on its $36T in debt are estimated to be $952 billion in 2025, up more than 175% in 5 years. Source: I/O Fund
The massive wall of debt that needs to be refinanced will likely be done now at much higher rates, adding even more to interest costs. For example, say that the $7 trillion in debt is refinanced at an average rate 1.5% to 2% higher, this would add an additional $105 to $140 billion annually in interest expenses simply from the higher rate structure.
Canadian mortgage lender First National says that “analysts reckon that every 30-basis point rise in the ten-year adds roughly $1.8 trillion to ten-year interest costs, sharpening the Treasury’s incentive to fund smoothly.” First National adds that 2025’s gross debt issuance will likely climb above $10 trillion based on the projected deficit and maturities, a volume that a modern market has not absorbed before.
Additionally, the U.S. dollar looks like it is heading lower, as measured by the dollar index (DXY). When this index is moving higher, money is flowing into U.S. markets, and when it is trending lower, there is a flight from U.S. markets. DXY still has, at least, one more drop in order to complete the downtrend pattern in play. This would target around $95 – $93, which is another 4% – 7% drop from current prices.
The US Dollar Index looks to have one more drop to $93-95 to complete its downtrend. Source: I/O Fund
This is a problem because the US needs foreign money flowing into its markets to finance our debt this year. For the first time since 2008, our total debt has meaningfully exceeded total domestic liquidity.
For the first time since 2008, the US’ total debt meaningfully exceeds total liquidity.
The U.S. alone simply does not have the needed liquidity to fund its debt, meaning that we must rely on foreign liquidity flows. Yet, as shown in DXY, foreign investments are fleeing the U.S. markets at the worst possible time.
Without foreign flows, rates have to rise until bonds find buyers – yet the predicament is that the US cannot afford rates to go higher as net interest payments then compound quicker.
A weaker dollar could also have numerous ramifications for the broader market. First, a weaker dollar could provide a tailwind to inflation as imports become more expensive, which in turn could force the Fed to keep rates higher for longer and prolong a rate cut cycle.
Second, approximately 25% of the outstanding debt, or around $9 trillion worth, is held by foreign investors – Japan with the largest holdings of more than $1.1 trillion, followed by the UK at ~$780 billion and China at $765 billion. Whereas higher rates tend to cause the dollar to strengthen by offering attractive returns for dollar-denominated investments, that’s not what we’re seeing after April’s trade policy announcements. From Morningstar:
“Conventional wisdom says new tariffs should have strengthened the dollar, since the import taxes were expected to reduce spending on goods produced overseas and shrink the trade deficit. A smaller trade deficit would mean the US would need to attract less foreign capital to keep the dollar from depreciating.”
Since the dollar is instead weakening, lower foreign appetite for debt could add more upward pressure to yields, and this fear resurfaced on Wednesday, as the weak 20-year auction pushed yields above expectations and sent equities sharply lower.
Subscribe for Full Access to the Article
It is easy to draw on one’s emotional bias and therefore build a believable case for what the market will do next. We think this is a mistake for investors positioning for the remainder of 2025.
Below the Advanced Tier Paywall is the Following information:
The specific game plan for how the I/O Fund plans to navigate the remainder of 2025 including the must-watch levels
The signals we are watching to gauge when the broad market tops and the exact levels where we will resume buying stocks.
Dial-in instructions for a 1-hour webinar on Thursday where I/O Fund Portfolio Manager, Knox Ridley, will discuss live the I/O Fund’s game plan for 2025. If you went into this sell-off fully invested without any risk management plan, we encourage you to attend our upcoming weekly webinar for premium members held this Thursday, May 29th at 4:30 ET.
Our five-year cumulative returns of 210% and 27.6% annualized places us at the top of Wall Street — if we were a hedge fund, we'd rank #2. If we were an ETF, we'd rank #5. We are celebrating our 5-year anniversary with our biggest sale of the year – take advantage with up to $275 off Advanced here. Check out the sale 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.
For our Premium Advanced Members, we discuss the following:
The specific game plan for how the I/O Fund plans to navigate the remainder of 2025 including the must-watch levels
The signals we are watching to gauge when the broad market tops and the exact levels where we will resume buying stocks.
Please keep an eye out for dial-in instructions for a 1-hour webinar on Thursday where I/O Fund Portfolio Manager, Knox Ridley, will discuss live the I/O Fund’s game plan for 2025. If you went into this sell-off fully invested without any risk management plan, we encourage our Advanced Members to attend our upcoming weekly webinar for premium members held this Thursday, May 29th at 4:30 ET.
Broad Market Analysis:
It is easy to draw on one’s emotional bias and therefore build a believable case for what the market will do next. We think this is a mistake for investors positioning for the remainder of 2025. Instead, we will continue to let the markets tell us what is to come.
While there is mounting evidence that the current bounce off the April lows could continue to new all-time highs, the unique risks being revealed in this market warrant caution. Furthermore, the larger pattern that has developed from the 2022 low is telling us that even if we do see a move to new highs, it will likely not be a prolonged trend higher before volatility picks back up.
“The next move will be a corrective rally that makes a lower high. The targets for this bounce are between 5600 – 6050.”
We further stated that once we see our first larger correction from this region, how the market corrects from there will likely determine the remainder of the year.
The below analysis outlines our specific game plan for how we plan to navigate the remainder of 2025…
As of now, the market has topped at 5968 and has the potential for one more small push higher. Regardless, we are in a topping pattern for the expected correction into the summer.
The I/O Fund’s S&P 500 outlook and game plan depending on how the next correction pans out. Source: I/O Fund
Red: In this scenario, we have are completing a rally that should make a lower high. This will set us up for a drop to new lows. The initial drop from the February 19th high was the A-wave. We are completing the lower high, B-wave, which will set us up for a 5-wave drop to new lows in the C wave. C-waves are always 5-wave patterns, so how we drop will be crucial for determining if this count is in play.
Green: This count would have us completing a larger correction within a bigger uptrend. If the coming drop is a messy and overlapping move that resembles a 3-wave pattern, it will likely make a higher low in an on-going bull market to new highs. This correction will target the 5600 region, first. As long as it holds 5100, we can maintain a setup to new highs around 6300 – 6500 in the coming months.
If we see a more direct drop that takes the shape of a 5-wave pattern, then the market is telling us the risks described in this report are likely greater than the market believes, as we set up for a drop below the April lows.
If on the other hand, we see a messy/overlapping 3-wave retrace that finds support in the 5600 – 5100 region, then it is the market telling us to look past these risks for now. This would be the set up for new all-time highs in the coming months.
I do want to state that even if we do see the scenario where we push to new all-time highs later in the year, the larger pattern in play suggests this will be the final 5th wave in the bull market that started at the 2022 lows. It should be accompanied by numerous key markets and stocks making higher lows. In other words, this would be a rally that we would likely sell into, as we set up for a more prolonged period of volatility.
Conclusion
The market breadth, Q1 earnings beats, and the size of this rally suggest that new all-time highs are likely to follow; however, one cannot underestimate the unique risks within the backdrop of markets. We have never seen more uncertainty in geo-political dynamics, which is forcing companies to withhold guidance as many of the Q1 beats are due to tariff pull forwards.
Furthermore, with nearly half of the U.S. government debt needing to be refinanced this year and next, the bond market continues to move lower in the face of market volatility and uncertainty. This is a trend we have not seen in over 30 years and could be signaling a sea change in how global markets interact moving forward.
If we see an aggressive 5 wave drop, we will position defensively for a move below the April low. If we see an overlapping, 3-wave move that holds over 5100, we will position for the 2nd best buying opportunity of 2025.
Join me Thursday, May 29th at 4:30 pm EST as I discuss this in further detail in a live webinar. The recorded version will be released later in the evening on the 29th.
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.
When you’re Babe Ruth, the crowd expects to see a homerun. Hit a single or a double and the fans go home disappointed. Nvidia has continued to report exceptional earnings results, yet Nvidia stock is competing with itself at this point.
Next week, Nvidia will report fiscal Q1 earnings, and the market has become accustomed to the company reporting a string of homeruns and grand slams. While Q1 results will be propped up by China stockpiling the H20s, the outlook for Q2 is the choppiest the company has faced in two years. This is because Nvidia has a narrower path than usual to impress investors as the Hopper generation demand is waning while Blackwell is (finally) shipping but not at the levels originally expected.
Last quarter, I published the analysis: “Nvidia Suppliers Send Mixed Signals for Delays on GB200 Systems – What it Means for Nvidia Stock.” which stated “Given market jitters around DeepSeek, which turned out to be a non-issue, something more material related to the GB200s, such as growth slowing below expectations at the start of the new fiscal year, could send the stock below $100 — which we would see as a buying opportunity […] ultimately, my firm trimmed our Nvidia position (to a 10% allocation) and will happily buy lower should the assumptions in this analysis materialize. Nvidia remains the stock of the decade; however, stock returns – and product launches — are not perfectly linear.” –February 2025
That analysis played out. We were able to buy Nvidia at $87, issuing a real-time trade alert that has returned 53% on that tranche since early April. We also added several key Nvidia suppliers that have moved sharply higher.
With a strong seven-year track record on this name, I felt it was important to share my perspective as Nvidia Week kicks off on Wall Street.
H20 Export Ban Will Result in $5.5B Inventory Loss and Cost $15B in Revenue
China has dominated the headlines over the past quarter and exactly how Nvidia plans to overcome geopolitical tensions will be a primary focus in the upcoming call. Export controls have been in place for years, hence the H800 and the H20 GPU variants, which were designed to be less powerful GPUs to comply with export restrictions. Yet, the license restrictions were changed in April, resulting in a $5.5 billion inventory loss for Nvidia. More recently, Jensen Huang clarified it would be $15 billion in revenue stating the inventory will have to be discarded.
Here is a brief summary of the USA-China GPU licensing restrictions:
In 2022, the Biden administration placed export controls on the A100s and H100s due to bandwidth, requiring the bandwidth to be lowered to 400GBps This led to the A800s and H800s.
In October of 2023, performance requirements were accounted for in the export controls, limiting sales of Nvidia’s A800, H800, L40, L40S and RTX 4090 chips. This led to Nvidia creating the H20 GPUs.
In April of 2025, the Trump administration has effectively banned the H20s by denying the export license. This is on the grounds that H20s can be used in supercomputers and offers 20% faster inference than the H100s.
While the H20 has reduced compute performance compared to the H100s — including fewer Tensor Cores and lower FP8/FP16 throughput — it retains high-speed interconnect capabilities through its support for NVLink and PCIe Gen4, and features 96GB of HBM3 memory with 4.0 TB/s of memory bandwidth. The 96GB of HBM3 exceeds the H100s, which is large enough for LLM models to run in memory and lower costs. The higher HBM3 also translates to the H20 offering fast communication when clustered with other GPUs for a multi-GPU system supercomputer.
Furthermore, although the H20 performs at roughly 50% of the H100s, it has a power advantage at 400 watts compared to the H100s 700 watts. Partially due to the lower power, while maintaining high-speed bandwidth for inference, means the H20s have remained attractive to Chinese firms especially in the wake of DeepSeek’s R-1 release.
Selling chips to China for use in supercomputers has been prohibited since 2022. Meanwhile, many industry experts believe Chinese firms were stockpiling the H20s to build a large supercomputer. The Institute for Progress, a nonpartisan group, wrote a long-form explanationof the loopholes that were being used, stating: “The United States is about to make another strategic mistake: Allowing three Chinese firms to receive over $16 billion in orders for NVIDIA H20 chips, amounting to over 1.3 million chips. This order is over six times the size of Colossus, the largest compute cluster in the world. It would more than double China’s entire existing stock of H20 chips. If these chips are delivered, they will dramatically increase Chinese firms’ ability to develop frontier AI models and deploy them at scale.”
As far as when the stockpiling began, semiconductor Insights analyst, Claus Aasholm, noted back in December that “The downgraded H20 system, which passes the embargo rules for China, is doing incredibly well. With 50% quarter-over-quarter growth, it is currently Nvidia’s most successful product. The H100 business “only” grew 25% QoQ.”
According to Reuters, analysts had forecast a total of $12 billion in H20 sales for Nvidia's fiscal year ending in January. However, China revenue for the year was significantly higher at $17.1 billion.
However, this pales in comparison to what Q1 and Q2 were about to report in terms of China revenue.
Nvidia Q1 Earnings Preview: Loss of China Revenue Will Sting
In Q1, Chinese tech companies such as Alibaba, ByteDance and Tencent were hurrying to place H20 orders. The Information reported that Chinese Big Tech companies had placed $16 billion worthof H20 chips in the first three months of the year.
This would represent a sudden surge of roughly 3X growth given previous quarters peaked at $5.5 billion:
Nvidia’s China Revenue:
Q1 2025 ending April 2024: $2.49 billion
Q2 2025 ending July 2024: $3.67 billion
Q3 2025 ending October 2024: $5.42 billion
Q4 FY25 ending Jan 2025: $5.52 billion
Pictured Above: Nvidia’s quarterly revenue in China was $5.4 billion for the October quarter and $5.52 in the January quarter (not pictured).for the October quarter and $5.52 in the January quarter (not pictured).
When Nvidia stated they would see a $5.5 billion inventory charge in Q1, it was suggesting a very high monthly run rate given the export restrictions were only in effect for the remaining three weeks of Q1. According to the SEC filing “First quarter results are expected to include up to approximately $5.5 billion of charges associated with H20 products for inventory, purchase commitments, and related reserves.”
If you view China revenue on a fiscal year basis, then The Information is suggesting that the first three months of the year resulted in nearly as much revenue from China as all of last year at $17.1 billion.
This helps to illustrate Nvidia was filling a lot of Chinese orders very suddenly a lot of Chinese orders very suddenly before the government intervened. This is further supported by the $16B figure from The Information as revenue of that magnitude from China is not seen in prior quarters.
Nvidia Q2 Earnings Guide Likely to be Impacted
Nvidia’s beats have become narrower over the past few quarters. Although it’s not clear what the impact will be in Q2, we have some indication from semiconductor peer AMD that the export ban of the MI308 will be mostly felt in Q2. It's logical to assume Nvidia will disclose something similar in their earnings call.
In the most recent quarter, Nvidia beat by $1.2 billion. At the start of the AI surge, Nvidia beat by $2.4 billion, and had initially raised guidance by $2 billion for a total upward surprise of $4.4 billion if you generously combine the May guidance with the August results.
If we look at fiscal year estimates of $200B and we take the $15B at face value — meaning there is no other impact further out into Q2 and Q3 (there very well could be if we assume $16B were rushed orders, yet the quarterly run rate was $5.5B, which is a floor for the other quarters) — then that’s 7.5% of revenue. This is not a reason to run for the hills, but it’s certainly revenue that has to be absorbed by other SKUs with strong potential Q1/Q2 is where the bulk of the impact is felt.
Analysts are preparing for lower QoQ growth with Q2 revenue $2.7B higher than Q1 compared to QoQ increase of $5.6 billion in Q3. Part of this is that Blackwell is (finally) ramping but not at the volume originally expected for this quarter.
Blackwell Revenue May Absorb China Losses
As the Blackwell vs Hopper GPU transition unfolds, investors are watching to see if Blackwell can make up for declining demand for the aging Hopper architecture as the H20 impact dissipates.
Nvidia said in Q4 that it delivered revenue of $11 billion for its Blackwell GPU, marking the fastest product ramp in its history. However, there was not a strong ramp in the GB200 NVL72 racks in Jan-March, with delivery estimates finally increasing in April.
Data from Morgan Stanley places Jan-March GB200 NVL72 rack shipments at ~1,000, while April shipments were estimated to rise sharply to ~1,500, with the majority coming from Foxconn. This aligns with fiscal Q2 GB200 NVL72 rack estimates of 4-5K, up 3-4x sequentially.
At the midpoint of those estimates, GB200 NVL72 rack revenue would calculate out to ~$13.5 billion at a $3 million ASP, complemented by HGX B200 shipments and other Blackwell products, which likely contributed 70%+ of the $11 billion from Q4.
This means even though Q2 could see a bigger impact from China that Blackwell sales may be able to help absorb these losses. The only hitch is that suppliers are not fully in agreement with that takeaway.
Subscribe for Full Access to the Article
Below the Paywall is the Following information
What AI-Related Suppliers and Competitors are Saying Ahead of Nvidia Earnings, including an important supplier correlation that broke down in November and worsened in April.
The timing of when Big Tech capex will hit peak growth and what this means for Nvidia’s stock
The I/O Fund’s trading plan for Nvidia including never-before published buy targets over a 12 to 18-month time frame.
IT’SOUR BIGGEST SALE OF THE YEAR — GET $275 OFF ADVANCEDFORFOURDAYSONLY. Expires May 27th at Midnight. 👉 Claim Offer Here
The last time we published our Nvidia price targets in Where I Plan to Buy Nvidia Stock Next, it was to say the stock would trade below $100. The stock topped five days later at $149. This buy plan was perfectly timed before the DeepSeek selloffs. We then reiterated this price target again the very week when Nvidia’s stock was at $138 and traded below $90 a little over a month later. Real-time trade alerts are sent to Members, including when we snagged shares as low as $87. When you layer these granular details on top of our first entry being $3.15 in 2018 — you will not want to miss the next buy plan our firm is putting into place.you will not want to miss the next buy plan our firm is putting into place.
Please note: The I/O Fund conducts research and draws conclusions for the company’s portfolio. We then share that information with our readers and offer real-time trade notifications. This is not a guarantee of a stock’s performance and it is not financial advice. Please consult your personal financial advisor before buying any stock in the companies mentioned in this analysis. Beth Kindig and the I/O Fund own shares in NVDA at the time of writing and may own stocks pictured in the charts.
For our Premium Members, we discuss the following:
What AI-Related Suppliers and Competitors are Saying Ahead of Nvidia Earnings, including an important supplier correlation that broke down in November and worsened in April.
The timing of when Big Tech capex will hit peak growth and what this means for Nvidia’s stock
The I/O Fund’s trading plan for Nvidia including never-before published buy targets over a 12 to 18-month time frame.
What AI-Related Suppliers and Competitors are Saying Ahead of Nvidia Earnings:
Unlike software, hardware ramps can take considerable time – in this case, Blackwell will have taken nearly 18 months from its announcement in March of 2024 until it ramps in volume in H2 2025. Given the improving, yet still muted commentary from Nvidia suppliers, the I/O Fund is looking toward Nvidia's August and November earnings calls as being the stronger earnings reports this year.
Here is what you need to know about Q1 and Q2 so far – some of which has been reported on while other notable points are being reported for the first time:
AMD to See $700M Impact in Q2; $1.5B Impact in FY2025
AMD is a solid company to track not only for its unique, underdog ascent in the AI market, but also because the company is similarly impacted with export license restrictions on the MI308 variant from tis MI300 series GPUs. Management stated the export restrictions will lead to a $700 million loss in revenue for AMD in Q2.
Here is a statement by CEO, Lisa Su:
“As a reminder, in April, a new export license requirement was put in place for MI308 shipments to China, the impact of which is included in our guidance. We expect revenue to be approximately $7.4 billion plus or minus $300 million. This includes an estimated $700 million revenue reduction as a result of the new export license requirement. Despite this headwind, the midpoint of our guidance represents 27% year-over-year revenue growth. For the full year 2025, we estimate the revenue impact due to the export license requirement to be approximately $1.5 billion.”
In terms of data center growth, it’s expected that AMD’s data center segment will decline in Q2 due to the loss of MI308 revenue yet will resume growth in Q3 and Q4.
Server Makers: Dell, Super Micro and HPE
When we look at AI server makers, there are signs that supply chain bottlenecks are easing — yet not at the pace originally expected for H1 2025.
Super Micro: the Nvidia Proxy
Super Micro is an obvious place to start when parsing out GPU shipments given the AI server maker led the market during the Hopper generation and grew from $7 billion in annual revenue to an expected $22 billion for the fiscal year ending in June – more than a 3X increase from the Hopper generation. When you look further out to include Ampere GPUs, revenue increased 7X from $3 billion in annual revenue to $22 billion in the current fiscal year ending in June.
Pictured Above: Supermicro’s revenue has correlated to Nvidia’s until only very recently. The disconnect that began in November has worsened in the latest quarter.
Super Micro offered a disappointing report as sales declined 19% QoQ and management lowered guidance for FY2025 revenue from $23.5B to $25B to $26B to $30B. At midpoint, this represents a 13.4% miss compared to previous guidance. Notably, this miss is concentrated in the current quarter as the company’s fiscal year ends in June.
While some are pointing toward a transition from Hopper to Blackwell as the issue, to be clear, it’s more likely the China impact from the H20 restrictions combined with delays for NVL72 volume shipments.
Regarding the NVL72s, Supermicro pointed toward direct liquid cooling as the primary hangup for the delays in the Q&A:
“Michael Ng
Great, thanks Charles, that's very helpful. And just as a follow up, can you talk about whether or not you're seeing differences in demand between HGX versus NVL 72 racks? Any differences there either in customer demand or your ability to fulfill demand on either product? Thank you.
Charles Liang
Yes, we see strong demand for kind of GB200 NVL 72 and B200 liquid cooling. But customer liquid cooling data center basically a little bit dead. So that's why they are waiting there, waiting a little bit more than what we expect. So but however the solution, their data center will be ready very soon and we to see our schedule is getting much more exciting now.”
Regarding the H20 impact, Supermicro’s Asia revenue increased 10 points QoQ, which suggests Supermicro was a beneficiary of an increase in China orders to some degree:
On an annual basis, Supermicro has not reported Asia revenue higher than 21.9% — proving the 29.4% in the latest quarter is certainly an outlier for this company.
Overall, both Supermicro and AMD point toward Nvidia’s Q2 as likely the choppiest quarter the company has faced in some time.
Additional Server Makers:
One data point does not make a trend, therefore, it’s prudent to check if these conclusions are being echoed elsewhere. As you’ll see below, the worst is likely behind us for Blackwell delays yet suppliers are not exactly surging in their AI segments ahead of deliveries in this specific quarter.
Dell:
Dell has not reported since February, which provides another month of visibility into how Q1 may fare yet does not offer much in terms of March or April. The company reported AI orders of $1.7 billion, down (53%) QoQ and shipments of $2.1 billion, down (28%) QoQ with $4.1 billion in backlog as customers work through “technology changes.”
Looking ahead, Dell foresees $15B in AI shipments this year yet the guide for the current quarter missed estimates. Dell’s quarter ends in April and the company guided Q1 revenue in the range of $22.5B to $23.5B, representing YoY growth of 3.4% at the midpoint, missing estimates by 3% yet expects adjusted EPS to grow 25% YoY to $1.65.
Dell is a mixed bag of course as there is significant consumer exposure on the PC side – yet interesting enough, analysts are more bullish on PCs outperforming in the upcoming quarter than AI servers due to a pull forward ahead of tariffs with Raymond James stating: ‘The AI transition between GPU generations has been more disruptive than anticipated, and checks suggest PC purchases have been pulled forward in anticipation of tariffs.”
Overall, we need to hear Dell’s update following Nvidia’s earnings before any conclusions can be drawn. What we do know is that in February, Dell was not too confident about their next quarter’s guide, hence lowering it by 3%.
HPE:
Hewlett-Packard is smaller in terms of AI revenue yet reported similar results as Dell in the March earnings report with AI systems mix falling 6.2% QoQ from 17.7% to 11.5%. According to the CEO, the headwinds are unlikely to clear up in Q2: “We recognized roughly $900 million of [AI systems] revenue, up from about $400 million last year, but down sequentially as expected due to chip availability and customer readiness. We expect these factors will continue to affect our AI systems business.” HPE has a mix of segments in total revenue, yet the company’s guidance missed Q2 estimates by 6.6%.
Looking forward, HPE’s statements match what others are saying, which is that Blackwell is finally ramping – although at lower levels than originally estimated: " In AI, we continue to see strong demand from model builders and service providers. We booked $1.6 billion in new AI system orders in the quarter, bringing our cumulative AI system orders to $8.3 billion. The Blackwell GPU generation of products represented approximately 70% of our new order intake in Q1.”
Vertiv:
Vertiv is not a server maker, rather provides the power supply and thermal management solutions required for data center infrastructure. The company lowered its guidance last quarter, yet raised guidance in the most recent quarter – pointing toward a successful resolution to thermal management issues for Blackwell rack-level systems. Here is what management stated: “The $150 million increase in organic sales is driven by both the first quarter and higher expectations in the second quarter versus what was implied in our prior guidance.”
On the call, an analyst asked if Vertiv “precedes the chip shipments” to which the CEO answered affirmatively by 3-6 months. Therefore, the encouraging inflection seen in Vertiv’s report is unlikely to result in an immediate correlation to Nvidia.
Note on Foxconn:
Foxconn (Hon Hai) is a crucial part of Nvidia's Blackwell supply chain, with it reportedly having the world’s largest GB200 manufacturing facility in Mexico. Foxconn said in Q1 that AI server revenue rose 50% YoY, and projected that Q2’s AI server revenue would double QoQ and YoY. Management explained that the reason Q1 did not double was “mainly due to the GB series entering mass production at the end of 1Q25,” and that most of those products would be delivered in 2Q25. Foxconn added that “HGX demand will continue to expand.”
For Q2, Foxconn said that AI servers were entering high-volume production, and would account for a larger portion of revenue at 50% of server sales, up from 40-42% in 2024. Foxconn also expects AI server revenue to improve each quarter of the year, and it reaffirmed guidance for AI server revenue to grow more than 50% YoY to surpass NT$1 trillion (US$33.0 billion) on high demand.
JP Morgan believes that Foxconn entered its large-scale ramp of GB200 production in late March, targeting 30,000 rack shipments for the full-year, with 10,000 of those being GB200/300 NVL72. This suggests that the ramp is still in the early stages, though accelerating into the summer months based on recent rack shipment estimates in April.
Obligatory Discussion on Capex
Capex and how it relates to AI spending needs no introduction at this point. Investors have never had it so easy as to track demand openly like we can with Big Tech’s disclosures on their quarterly and fiscal year capex guidance. Here is an overview of just how current guidance from Big Tech:
Amazon has forecast capex of more than $100 billion this year, up from $78 billion in 2024. This represents the largest amount being spent by a single tech company and is a higher mix of custom silicon compared to GPUs.
Alphabet has forecast capex of $75 billion this year, up from $52.5 billion in 2024.
Meta has forecast capex of $68 billion this year, up from $39.2 billion in 2024. This represents the largest growth among the Big Tech companies this calendar year at 74%.
Microsoft has forecast capex of $80 billion this fiscal year ending in June, up from $44.5 billion. Given Microsoft has a mid-year fiscal year, this represents the largest growth among tech companies over the past 12 months at 80%.
I want to caution that peak years for AI capex growth are likely behind us. Collectively, Microsoft, Amazon, Meta and Google are projected to spend more than $330B on capex in 2025, up nearly 34% YoY. According to UBS, estimates for capex will rise less than 10% YoY to $364B, with Amazon’s capex nearly flat and Meta seeing the largest increase at 15% YoY.
Wall Street has consistently placed capex estimates too low, meaning it’s likely we see greater than a 10% YoY increase next year, yet the chances growth rates accelerate beyond this year’s 34% growth is not likely (hence the statement we’ve likely hit peak growth).
Microsoft recently offered a glimpse that the voracious appetite to grow AI infrastructure may eventually come back down to earth. Management said capex in fiscal 2026 (beginning in the second half of calendar 2025) will grow at a slower pace than FY2025, with a higher mix of short-lived assets. Additionally, capex declined sequentially for the first time in 2 years, at $21.4 billion versus $22.6 billion in the prior quarter. Q3’s figure was also slightly lower than expected due to variability in timing of data center leases, though capex is expected to increase sequentially in fiscal Q4.
There are many avenues for Nvidia to continue its AI dominance, which we’ve covered for the past 2-3 years on the premium site to prepare our readers for Nvidia's long runway. Flat capex will likely spook the markets (more likely in January 2026 than in July) yet we would be buyers as we are quite clear Nvidia’s AI thesis goes well beyond AI servers and infrastructure.
I/O Fund’s Buy Plan
By Knox Ridley
Since launching the I/O Fund live portfolio in May of 2020, Nvidia is one of only three positions we have owned without interruption. Since 2021, Nvidia has remained in the top three, which was two years ahead of the AI surge.
On a more granular level, we backed up our research with 9 buy alerts that we sent to our Members when NVDA was under $20 from 2021 – 2022. Being early to the A.I. trend has been one of the primary reasons why the I/O Fund was able to outperform our benchmarks and competition in 2023 – 2025.
However, unlike many, we favor an active approach – i.e., taking gains in positions, and reducing risk, regardless of how bullish our long-term thesis may be. Stocks do not move in a straight line, and although we believe that we are in the early innings of A.I., we think investors should expect periods of volatility that punctuate the larger trend higher.
For this reason, we cut ¼ of our Nvidia position in June of 2024 at $129.49, just before Nvidia saw a greater than 35% drawdown into August. We further cut half of our position at $127 and $140 on February 6th and 20th of this year, just before Nvidia dropped more than 40% into the April 7th lows.
While both technical and fundamental concerns were the reasons behind these decisions to take gains, we were able to follow up these sales with buys at much lower prices. In our January 2025 article, titled, “Where I Plan to Buy Nvidia Stock Next,”we outlined our long-term targets in the $90s and $80s. We were able to execute this plan on April 4th at $94 and at $87 on April 7th. These were prices targets that I had been discussing with our premium members in my weekly Thursday webinar for many months.
Now that we have seen 54% bounce off the April lows, we believe Nvidia, and the markets are at a major inflection point.
Technical Analysis
We’ll begin with the bigger trend that started on the October lows in 2022. You can see how vertical the pattern is. This is a standard 5-wave pattern, which can allow for two general interpretations on where Nvidia is likely heading next:
Blue – This is our primary analysis and suggests that the April low was the end of the larger 4th wave decline. This would mean that Nvidia is setting up for a 5th wave push to new highs, which should target $170 – $195 then $240 – $295. This will complete the larger uptrend pattern and likely give way to a multi-month correction.
Red – This scenario has the August 2024 low as the bottom of the 4th wave, and the weak and messy push higher into the February high as the final 5th wave in this uptrend pattern. The next larger drop should be a more direct pattern that ultimately breaks below $97. This will setup the final drop to the $70s – $60 region.
NVDA is setting up for a reversal. Note the RSI indicator below. It is currently at the same level that the February top occurred, as well as the first major drawdown in late 2022. This is a key region where bounces tend to fail. Also, note how volume continues to drop the higher we go. There are less buyers the higher we go, which is not a good sign.
What will be key is how Nvidia retraces. If we see a messy and overlapping drop that resembles a 3-wave drop, it is signaling that the bullish blue count is likely in play. On the other hand, if we see a more aggressive and direct drop that resembles 5-wave drop, it will support the probability that the red count is what is in play.
If we zoom in on the current 2025 trend, we can get a better idea of what these paths might look like. This is also the chart we will use to establish a risk management plan for any new buys.
Nvidia broke out from the down trend line that started at the February 2025 top. This is historically a bullish signal, and one that favors the bullish scenario. Furthermore, the smaller degree count looks incomplete. As long as $129 holds, I expect NVDA to push to the $143 – $149 region before putting in a local top.
As stated before, if the next drop is a 3-wave move that is messy and overlapping, we will target the $116, $110, $103 region for an additional buy. As long as this drop holds over $97, we expect to see a low, followed by a larger push into the $240 – $294 region in the coming months. If we instead see a more direct drop that resembles a 5-wave drop, we will be on high alert. A break below $97 will setup a final drop into the $76 – $60 region. This will setup a tremendous buying opportunity, if it happens.
The I/O Fund is closely monitoring Nvidia for a potential entry point. Join us Thursdays at 4:30 p.m. in our Advanced Market webinars, where we’ll outline our strategy for initiating a position with maximum upside in mind. Learn more here.
Pro Members: Don’t miss our biggest sale of the year — save $275 on an annual Advanced Market Signals plan. Email us to upgrade
Please note: The I/O Fund conducts research and draws conclusions for the company’s portfolio. We then share that information with our readers and offer real-time trade notifications. This is not a guarantee of a stock’s performance and it is not financial advice. Please consult your personal financial advisor before buying any stock in the companies mentioned in this analysis. Beth Kindig and the I/O Fund own shares in NVDA at the time of writing and may own stocks pictured in the charts.
For our Premium Members, we discuss the following:
What AI-Related Suppliers and Competitors are Saying Ahead of Nvidia Earnings, including an important supplier correlation that broke down in November and worsened in April.
The timing of when Big Tech capex will hit peak growth and what this means for Nvidia’s stock
The I/O Fund’s trading plan for Nvidia including never-before published buy targets over a 12 to 18-month time frame.
What AI-Related Suppliers and Competitors are Saying Ahead of Nvidia Earnings:
Unlike software, hardware ramps can take considerable time – in this case, Blackwell will have taken nearly 18 months from its announcement in March of 2024 until it ramps in volume in H2 2025. Given the improving, yet still muted commentary from Nvidia suppliers, the I/O Fund is looking toward Nvidia's August and November earnings calls as being the stronger earnings reports this year.
Here is what you need to know about Q1 and Q2 so far – some of which has been reported on while other notable points are being reported for the first time:
AMD to See $700M Impact in Q2; $1.5B Impact in FY2025
AMD is a solid company to track not only for its unique, underdog ascent in the AI market, but also because the company is similarly impacted with export license restrictions on the MI308 variant from tis MI300 series GPUs. Management stated the export restrictions will lead to a $700 million loss in revenue for AMD in Q2.
Here is a statement by CEO, Lisa Su:
“As a reminder, in April, a new export license requirement was put in place for MI308 shipments to China, the impact of which is included in our guidance. We expect revenue to be approximately $7.4 billion plus or minus $300 million. This includes an estimated $700 million revenue reduction as a result of the new export license requirement. Despite this headwind, the midpoint of our guidance represents 27% year-over-year revenue growth. For the full year 2025, we estimate the revenue impact due to the export license requirement to be approximately $1.5 billion.”
In terms of data center growth, it’s expected that AMD’s data center segment will decline in Q2 due to the loss of MI308 revenue yet will resume growth in Q3 and Q4.
Server Makers: Dell, Super Micro and HPE
When we look at AI server makers, there are signs that supply chain bottlenecks are easing — yet not at the pace originally expected for H1 2025.
Super Micro: the Nvidia Proxy
Super Micro is an obvious place to start when parsing out GPU shipments given the AI server maker led the market during the Hopper generation and grew from $7 billion in annual revenue to an expected $22 billion for the fiscal year ending in June – more than a 3X increase from the Hopper generation. When you look further out to include Ampere GPUs, revenue increased 7X from $3 billion in annual revenue to $22 billion in the current fiscal year ending in June.
Pictured Above: Supermicro’s revenue has correlated to Nvidia’s until only very recently. The disconnect that began in November has worsened in the latest quarter.
Super Micro offered a disappointing report as sales declined 19% QoQ and management lowered guidance for FY2025 revenue from $23.5B to $25B to $26B to $30B. At midpoint, this represents a 13.4% miss compared to previous guidance. Notably, this miss is concentrated in the current quarter as the company’s fiscal year ends in June.
While some are pointing toward a transition from Hopper to Blackwell as the issue, to be clear, it’s more likely the China impact from the H20 restrictions combined with delays for NVL72 volume shipments.
Regarding the NVL72s, Supermicro pointed toward direct liquid cooling as the primary hangup for the delays in the Q&A:
“Michael Ng
Great, thanks Charles, that's very helpful. And just as a follow up, can you talk about whether or not you're seeing differences in demand between HGX versus NVL 72 racks? Any differences there either in customer demand or your ability to fulfill demand on either product? Thank you.
Charles Liang
Yes, we see strong demand for kind of GB200 NVL 72 and B200 liquid cooling. But customer liquid cooling data center basically a little bit dead. So that's why they are waiting there, waiting a little bit more than what we expect. So but however the solution, their data center will be ready very soon and we to see our schedule is getting much more exciting now.”
Regarding the H20 impact, Supermicro’s Asia revenue increased 10 points QoQ, which suggests Supermicro was a beneficiary of an increase in China orders to some degree:
On an annual basis, Supermicro has not reported Asia revenue higher than 21.9% — proving the 29.4% in the latest quarter is certainly an outlier for this company.
Overall, both Supermicro and AMD point toward Nvidia’s Q2 as likely the choppiest quarter the company has faced in some time.
Additional Server Makers:
One data point does not make a trend, therefore, it’s prudent to check if these conclusions are being echoed elsewhere. As you’ll see below, the worst is likely behind us for Blackwell delays yet suppliers are not exactly surging in their AI segments ahead of deliveries in this specific quarter.
Dell:
Dell has not reported since February, which provides another month of visibility into how Q1 may fare yet does not offer much in terms of March or April. The company reported AI orders of $1.7 billion, down (53%) QoQ and shipments of $2.1 billion, down (28%) QoQ with $4.1 billion in backlog as customers work through “technology changes.”
Looking ahead, Dell foresees $15B in AI shipments this year yet the guide for the current quarter missed estimates. Dell’s quarter ends in April and the company guided Q1 revenue in the range of $22.5B to $23.5B, representing YoY growth of 3.4% at the midpoint, missing estimates by 3% yet expects adjusted EPS to grow 25% YoY to $1.65.
Dell is a mixed bag of course as there is significant consumer exposure on the PC side – yet interesting enough, analysts are more bullish on PCs outperforming in the upcoming quarter than AI servers due to a pull forward ahead of tariffs with Raymond James stating: ‘The AI transition between GPU generations has been more disruptive than anticipated, and checks suggest PC purchases have been pulled forward in anticipation of tariffs.”
Overall, we need to hear Dell’s update following Nvidia’s earnings before any conclusions can be drawn. What we do know is that in February, Dell was not too confident about their next quarter’s guide, hence lowering it by 3%.
HPE:
Hewlett-Packard is smaller in terms of AI revenue yet reported similar results as Dell in the March earnings report with AI systems mix falling 6.2% QoQ from 17.7% to 11.5%. According to the CEO, the headwinds are unlikely to clear up in Q2: “We recognized roughly $900 million of [AI systems] revenue, up from about $400 million last year, but down sequentially as expected due to chip availability and customer readiness. We expect these factors will continue to affect our AI systems business.” HPE has a mix of segments in total revenue, yet the company’s guidance missed Q2 estimates by 6.6%.
Looking forward, HPE’s statements match what others are saying, which is that Blackwell is finally ramping – although at lower levels than originally estimated: " In AI, we continue to see strong demand from model builders and service providers. We booked $1.6 billion in new AI system orders in the quarter, bringing our cumulative AI system orders to $8.3 billion. The Blackwell GPU generation of products represented approximately 70% of our new order intake in Q1.”
Vertiv:
Vertiv is not a server maker, rather provides the power supply and thermal management solutions required for data center infrastructure. The company lowered its guidance last quarter, yet raised guidance in the most recent quarter – pointing toward a successful resolution to thermal management issues for Blackwell rack-level systems. Here is what management stated: “The $150 million increase in organic sales is driven by both the first quarter and higher expectations in the second quarter versus what was implied in our prior guidance.”
On the call, an analyst asked if Vertiv “precedes the chip shipments” to which the CEO answered affirmatively by 3-6 months. Therefore, the encouraging inflection seen in Vertiv’s report is unlikely to result in an immediate correlation to Nvidia.
Note on Foxconn:
Foxconn (Hon Hai) is a crucial part of Nvidia's Blackwell supply chain, with it reportedly having the world’s largest GB200 manufacturing facility in Mexico. Foxconn said in Q1 that AI server revenue rose 50% YoY, and projected that Q2’s AI server revenue would double QoQ and YoY. Management explained that the reason Q1 did not double was “mainly due to the GB series entering mass production at the end of 1Q25,” and that most of those products would be delivered in 2Q25. Foxconn added that “HGX demand will continue to expand.”
For Q2, Foxconn said that AI servers were entering high-volume production, and would account for a larger portion of revenue at 50% of server sales, up from 40-42% in 2024. Foxconn also expects AI server revenue to improve each quarter of the year, and it reaffirmed guidance for AI server revenue to grow more than 50% YoY to surpass NT$1 trillion (US$33.0 billion) on high demand.
JP Morgan believes that Foxconn entered its large-scale ramp of GB200 production in late March, targeting 30,000 rack shipments for the full-year, with 10,000 of those being GB200/300 NVL72. This suggests that the ramp is still in the early stages, though accelerating into the summer months based on recent rack shipment estimates in April.
Obligatory Discussion on Capex
Capex and how it relates to AI spending needs no introduction at this point. Investors have never had it so easy as to track demand openly like we can with Big Tech’s disclosures on their quarterly and fiscal year capex guidance. Here is an overview of just how current guidance from Big Tech:
Amazon has forecast capex of more than $100 billion this year, up from $78 billion in 2024. This represents the largest amount being spent by a single tech company and is a higher mix of custom silicon compared to GPUs.
Alphabet has forecast capex of $75 billion this year, up from $52.5 billion in 2024.
Meta has forecast capex of $68 billion this year, up from $39.2 billion in 2024. This represents the largest growth among the Big Tech companies this calendar year at 74%.
Microsoft has forecast capex of $80 billion this fiscal year ending in June, up from $44.5 billion. Given Microsoft has a mid-year fiscal year, this represents the largest growth among tech companies over the past 12 months at 80%.
I want to caution that peak years for AI capex growth are likely behind us. Collectively, Microsoft, Amazon, Meta and Google are projected to spend more than $330B on capex in 2025, up nearly 34% YoY. According to UBS, estimates for capex will rise less than 10% YoY to $364B, with Amazon’s capex nearly flat and Meta seeing the largest increase at 15% YoY.
Wall Street has consistently placed capex estimates too low, meaning it’s likely we see greater than a 10% YoY increase next year, yet the chances growth rates accelerate beyond this year’s 34% growth is not likely (hence the statement we’ve likely hit peak growth).
Microsoft recently offered a glimpse that the voracious appetite to grow AI infrastructure may eventually come back down to earth. Management said capex in fiscal 2026 (beginning in the second half of calendar 2025) will grow at a slower pace than FY2025, with a higher mix of short-lived assets. Additionally, capex declined sequentially for the first time in 2 years, at $21.4 billion versus $22.6 billion in the prior quarter. Q3’s figure was also slightly lower than expected due to variability in timing of data center leases, though capex is expected to increase sequentially in fiscal Q4.
There are many avenues for Nvidia to continue its AI dominance, which we’ve covered for the past 2-3 years on the premium site to prepare our readers for Nvidia's long runway. Flat capex will likely spook the markets (more likely in January 2026 than in July) yet we would be buyers as we are quite clear Nvidia’s AI thesis goes well beyond AI servers and infrastructure.
I/O Fund’s Buy Plan
By Knox Ridley
Since launching the I/O Fund live portfolio in May of 2020, Nvidia is one of only three positions we have owned without interruption. Since 2021, Nvidia has remained in the top three, which was two years ahead of the AI surge.
On a more granular level, we backed up our research with 9 buy alerts that we sent to our Members when NVDA was under $20 from 2021 – 2022. Being early to the A.I. trend has been one of the primary reasons why the I/O Fund was able to outperform our benchmarks and competition in 2023 – 2025.
However, unlike many, we favor an active approach – i.e., taking gains in positions, and reducing risk, regardless of how bullish our long-term thesis may be. Stocks do not move in a straight line, and although we believe that we are in the early innings of A.I., we think investors should expect periods of volatility that punctuate the larger trend higher.
For this reason, we cut ¼ of our Nvidia position in June of 2024 at $129.49, just before Nvidia saw a greater than 35% drawdown into August. We further cut half of our position at $127 and $140 on February 6th and 20th of this year, just before Nvidia dropped more than 40% into the April 7th lows.
While both technical and fundamental concerns were the reasons behind these decisions to take gains, we were able to follow up these sales with buys at much lower prices. In our January 2025 article, titled, “Where I Plan to Buy Nvidia Stock Next,”we outlined our long-term targets in the $90s and $80s. We were able to execute this plan on April 4th at $94 and at $87 on April 7th. These were prices targets that I had been discussing with our premium members in my weekly Thursday webinar for many months.
Now that we have seen 54% bounce off the April lows, we believe Nvidia, and the markets are at a major inflection point.
Technical Analysis
We’ll begin with the bigger trend that started on the October lows in 2022. You can see how vertical the pattern is. This is a standard 5-wave pattern, which can allow for two general interpretations on where Nvidia is likely heading next:
Blue – This is our primary analysis and suggests that the April low was the end of the larger 4th wave decline. This would mean that Nvidia is setting up for a 5th wave push to new highs, which should target $170 – $195 then $240 – $295. This will complete the larger uptrend pattern and likely give way to a multi-month correction.
Red – This scenario has the August 2024 low as the bottom of the 4th wave, and the weak and messy push higher into the February high as the final 5th wave in this uptrend pattern. The next larger drop should be a more direct pattern that ultimately breaks below $97. This will setup the final drop to the $70s – $60 region.
NVDA is setting up for a reversal. Note the RSI indicator below. It is currently at the same level that the February top occurred, as well as the first major drawdown in late 2022. This is a key region where bounces tend to fail. Also, note how volume continues to drop the higher we go. There are less buyers the higher we go, which is not a good sign.
What will be key is how Nvidia retraces. If we see a messy and overlapping drop that resembles a 3-wave drop, it is signaling that the bullish blue count is likely in play. On the other hand, if we see a more aggressive and direct drop that resembles 5-wave drop, it will support the probability that the red count is what is in play.
If we zoom in on the current 2025 trend, we can get a better idea of what these paths might look like. This is also the chart we will use to establish a risk management plan for any new buys.
Nvidia broke out from the down trend line that started at the February 2025 top. This is historically a bullish signal, and one that favors the bullish scenario. Furthermore, the smaller degree count looks incomplete. As long as $129 holds, I expect NVDA to push to the $143 – $149 region before putting in a local top.
As stated before, if the next drop is a 3-wave move that is messy and overlapping, we will target the $116, $110, $103 region for an additional buy. As long as this drop holds over $97, we expect to see a low, followed by a larger push into the $240 – $294 region in the coming months. If we instead see a more direct drop that resembles a 5-wave drop, we will be on high alert. A break below $97 will setup a final drop into the $76 – $60 region. This will setup a tremendous buying opportunity, if it happens.
The I/O Fund is closely monitoring Nvidia for a potential entry point. Join us Thursdays at 4:30 p.m. in our Advanced Market webinars, where we’ll outline our strategy for initiating a position with maximum upside in mind. Learn more here.
Essentials Members: Don’t miss our biggest sale of the year — save $275 on an annual Advanced Market Signals plan. Email us to upgrade
Please note: The I/O Fund conducts research and draws conclusions for the company’s portfolio. We then share that information with our readers and offer real-time trade notifications. This is not a guarantee of a stock’s performance and it is not financial advice. Please consult your personal financial advisor before buying any stock in the companies mentioned in this analysis. Beth Kindig and the I/O Fund own shares in NVDA at the time of writing and may own stocks pictured in the charts.
AppLovin easily topped revenue and EPS estimates in Q1, but more importantly, the company is setting up for an additional under-reported catalyst with its web-based ad platform expected to launch its self-serve feature and scale with a wider pool of advertisers as the year progresses.
In addition, the company divested its App segment, which is the gaming assets portfolio, and is now a pureplay ad-tech stock. The high-growth and high-margin advertising business that ignited AppLovin’s strong returns over the past few years is now the company’s sole focus.
You’d be hard pressed to find a stronger stock in terms of fundamentals on the market today. There is plenty of runway left for this stock should the growth of 30%+ coupled with 80%+ gross margins and nearly 40% net margin continue. Consider that EPS grew triple digits this quarter and FY2026 EPS estimates are being revised higher by an astonishing $3.50 in incremental EPS.
In analysis below, we turn our focus to ways the company can sustain this growth on the top line and bottom line as we look at 2025 and beyond.
Ad Growth of 71% YoY
AppLovin reported 40.3% YoY revenue growth to $1.48 billion in the first quarter, beating consensus estimates by $100 million. This was AppLovin’s sixth consecutive quarter with revenue growth >35% YoY.
Advertising revenue increased 70.9% YoY to $1.16 billion, slowing slightly from 91% in the year-ago quarter. Management said growth was driven by continued enhancements in its AI ad engine, as well as the full quarter impact of its web-based ad solution even coming off the seasonally high e-commerce quarter in Q4.
For Q2, management guided Advertising revenue of $1.195 to $1.215 billion, pointing to 69.5% YoY growth at midpoint, maintaining its hypergrowth phase.
Management is still looking to explore CTV as a future growth channel, following its recent push into e-commerce, while the upcoming launch of its AI-powered self-service campaign management platform is expected to be both a catalyst for revenue and margins.
Given the Apps business is being divested, AppLovin will be reporting headline growth in the 60% range that is aligned with its Ads business rather than a mix of both. Consensus revenue growth estimates are much lower and show a sharp deceleration, as these comps still take into account revenue from the Apps segment . Thus, growth rates such as 20% in Q3 do not reflect the true performance of the business.
Apps revenue declined (14.4%) YoY to $325 million. AppLovin announced that it entered a definitive agreement to sell the segment to Tripledot Studios for $400 million in cash ($150 million at closing and a $250 million promissory note) and a 20% stake in Tripledot’s equity. The transaction is expected to close in Q2. This will transition AppLovin into an advertising pure-play.
Margins Show Continued Strength
Though AppLovin’s top-line growth is quite impressive, margins are where it shines, with gross margin surpassing 80% and operating margin reaching a new high. This combination of strong revenue growth and strong margins is driving exceptional operating leverage with triple-digit earnings growth.
Per management on the call: “Total revenue soared 40% from the same period last year to $1.5 billion, and adjusted EBITDA increased a remarkable 83% to an impressive $1 billion, achieving a fantastic 68% adjusted EBITDA margin […] Shifting to the Advertising business, we generated $1.16 billion in revenue and $943 million in adjusted EBITDA, achieving an incredible 81% margin.”
Gross margin expanded 5 points sequentially and more than 9 points YoY to 81.7%. Notably, AppLovin cut its cost of revenue by nearly (9%) YoY, from $294.1 million to $272.2 million, while still driving 40% total revenue growth and 70% advertising growth.
Operating margin remained above 44% for a third straight quarter at 44.7%. To put in perspective how strong these margins are, AppLovin would have a Rule of 40 score of 85% based on Palantir’s definition of revenue growth + operating margin, while Palantir had a score of 83%.
Post-divestment, AppLovin’s operating margin will look much different, given that Advertising’s current operating margin is likely much closer to 70%. Management talked about the Rule of 150 on the earnings call, reflecting the new fundamental structure of the company, with ~70% revenue growth and 70% operating margins:
“I don't know of any other tech company with the financial profile that we have and scale growing the way we are. I think it's on a Rule of 150 or something. And what we're focused on when we talk about priorities is how's 2026 going to be? How's '27 going to be?”
Net margin in Q1 was 38.8%, up more than 16 points YoY. AppLovin’s business model sees a high percentage of its operating income flow through to the bottom line, driving tremendous EPS growth as margins expand.
Adjusted EBITDA margin was 68%, well ahead of guidance for 63% to 64%, as adjusted EBITDA surpassed $1 billion. Advertising adjusted EBITDA margin expanded 8 points YoY and 3 points QoQ to 81%, with adjusted EBITDA of $943 million coming in nearly 16% ahead of guidance. Apps adjusted EBITDA margin was 19%, flat QoQ but up 4 points YoY.
For Q2, management guided for adjusted EBITDA of $970-990 million for an 81% margin.
EPS Grows Triple-Digits in Q1
AppLovin reported massive 149% YoY growth in GAAP EPS to $1.67, outpacing revenue growth by more than 3x. Looking ahead, EPS estimates have been revised significantly higher since our latest update in February, AppLovin: Expanding from Gaming to E-Commerce (and Beyond)AppLovin: Expanding from Gaming to E-Commerce (and Beyond).
Q2 EPS is now seen growing 125% YoY to $2.00, before rising to $2.16 in Q3 and exiting the year at $2.46.
This compares to February’s estimates for 60% growth to $1.43 in Q2, $1.66 in Q3 and $1.78 in Q4. This is a significant >38% increase for Q2 and Q4’s EPS, as AppLovin will benefit from a much leaner business model as an ad-tech pure play, with operating margins set to expand post-divestment to nearly 70%, aided by prudent cost management – sales & marketing expenses were down nearly (20%) YoY, R&D was down nearly (21%), and data center costs rose by just $30 million YoY on a $480 million increase in revenue.
For FY25, analysts now estimate AppLovin will generate $7.80 in EPS, up 72.3% YoY, with FY26 EPS rising 42% to $11.80. This is more than a $3.50 increase for FY26 since February’s $8.27 estimate.
Cash Flows and Balance Sheet
AppLovin’s cash flows are exceptional, with operating and free cash flow margins expanding to new records in Q1. Per the opening remarks: “In the first quarter, we generated $826 million in free cash flow, up a staggering 113% year-over-year. Quarter-over-quarter, our free cash flow grew 19%, representing an impressive 82% flow-through from adjusted EBITDA to free cash flow.”
Operating cash flow rose 112% YoY to $831.7 million for a 56% margin, expanding from a 51.1% margin in Q4 and nearly 19 points higher than 37.1% in the year ago quarter.
Free cash flow rose 113% YoY to $825.7 million for a 55.6% margin, expanding from 50.6% in Q4 and 36.6% in the year ago quarter.
Cash and cash equivalents decreased to $551 million from $741.4 million in Q4, though the closing of the Apps sale should help restrengthen its cash position. AppLovin also repurchased and withheld 3.4 million shares in Q1 for a total cost of $1.2 billion, funded primarily via free cash flow and a temporary draw on its revolving credit facility, which was already repaid.
Debt totaled $3.71 billion, and debt-to-equity ratio has surged from 3.4x last quarter to nearly 6.5x now, as AppLovin is now more highly levered due to the decrease in cash and increase in total liabilities from $4.78 billion to $5.13 billion.
Analyst Estimates Do Not Represent the Full Picture
Analyst estimates show very low growth because these are taking the combined business of gaming and ads, and basing growth on this whereas organic growth will be much higher.
While Q2 will still have five weeks’ revenue impact from gaming, Q3 and Q4 will be AppLovin’s first two quarters post-divestment, with consensus revenue growth of 20-23%.
However, organic growth for Advertising was guided at just under 70% in Q2, and expected to be 65% YoY in Q3 based on the current estimate for $1.38 billion in revenue, which aligns with trends for 15% QoQ growth from Q2’s guide. Revenue is expected to decelerate to 52% in Q4, though this comes against a 73% growth comp.
Looking at just the Advertising business, full-year revenue growth is projected at ~63.2% YoY to $5.26 billion, based on current guidance and estimates for 2H. This would value AppLovin at a rather pricey 23x PS for 2025, and if revenue grows 30% YoY in 2026 to ~$6.83 billion, AppLovin would be valued at 17.7x forward PS.
If AppLovin can lever strong execution, expansions into web-advertising and further into e-commerce, and better optimizations to its AI ad engine to drive 50% YoY growth in 2026, revenue would project out to $7.89 billion, or ~15.3x forward PS
On the bottom line, AppLovin is currently operating near a 40% net margin with some quarterly fluctuations, and Advertising’s strong operating margin profile will likely pull net margins to the high-50% to 60% range. Looking out to 2026, AppLovin could generate earnings of $11.85 on the 30% growth forecast, assuming a 58% net margin with ~10 million share buybacks. This would value it slightly above 30x forward PE. Based on the 50% growth forecast and similar net margin and buyback assumptions, earnings would project to $13.70, or growth of >72% YoY based on current FY25 estimates, valuing AppLovin at 26.1x forward PE.
Growth Catalysts for 2025 and Beyond:
In terms of executing a successful pivot, Applovin’s management team does not get enough credit. Mobile games are a business that has plateaued (that’s putting it nicely – it’s actually a market that has tanked). Applovin aggressively acquired mobile games and leveraged their mobile IP portfolio to build a formidable database of 1.4 billion users. By building an ad-tech business and acquiring AI engineering talent, the company was early to AI with its AXON 2.0 platform. The pivot is one of the boldest I’ve seen in a 15-year career in tech; on par with Lisa Su’s move to overtake Intel.
“The ad engine AXON 2.0 offers a monumental advantage to AppLovin as the company is ahead in the race for AI-driven advertising. AppLovin is also an arbitrage advertising platform, which means they can quantify the impact of their reach for advertisers by returning back to the advertiser what was spent or more within 30 days. If an advertiser spends $10,000 (or multiples of this), AppLovin is able to return that or more to the advertiser. The company is also unique in that it offers performance marketing for brands and direct-to-consumer. The Trade Desk primarily works with agencies, whereas AppLovin is attracting smaller and medium sized businesses that rely on performance. Most importantly, AXON 2.0 is an AI-powered advertising engine that is continuously improving. Every quarter and every year, AXON becomes more effective by ingesting more data that improves the model through self-learning.”
Looking into the future, however, the management team cannot rest on their laurels as gaming eventually hits its limit in inventory. Although 1.4 billion daily active users is impressive, at about half of what Meta has with its family of apps at 3 billion users, one could argue that gamers are only worth so much to a marketer as the demographic is narrower and more limited.
Applovin’s next moves are the following:
Branch out to e-commerce — This plays nicely into the restrictions AppLovin has with a cookie window to convert within 24 hours. Meaning, if a user converts beyond 24 hours, it is hard for AppLovin to verify attribution. Therefore, the company is less appealing to an auto advertiser where the buying decision is quite long compared to a T-shirt company.
Web-based advertising – mobile games are an app-based business, hence the name AppLovin. The company came to market in the mobile era, yet the company is not capitalizing on websites at this time.
Self-service Platform – although this goes hand-in-hand with the web-based advertising catalyst, it’s important to look at this feature separately as the onboarding of advertisers can scale more quickly once this feature launches. Given AppLovin’s ideal advertisers convert within 24 hours for products that are less than $250, self-serve platform is the only way forward that makes sense.
Go Global – this is not on the near-term product road map yet is a lever AppLovin can pull when the timing is right. The company is focused on the United States market, which is by far the most lucrative.
E-commerce Apps:
According to the CEO’s response to short sellers in February, the initial launch of the e-commerce platform has seen 600 advertisers with an annual run rate of $1 billion. In terms of increasing TAM (which is also related to the information below in the web-based advertising section), the company stated they are “sub-0.1% penetration in the market,” signifying a long runway.
Where the TAM is a bit constrained is Applovin has 24 hours on the attribution side to convert and this tends to perform best with products priced under $250.
“On web, we built the product to be self-attributing, so our own attribution platform. And it's not high turnover products. I mean, like, most products in the world are not selling something greater than $250. Our product — our models can go deliver something that's a couple hundred dollars within a few minutes of the ad being seen, and it it's happening quite often. I mean, obviously, scaled at the $1 billion run rate that I mentioned.”
The breakdown of mobile versus web-based advertising in terms of ad spend is as follows.
Total Ad Spend in the United States: $309B per year
In-app advertising represents $165.9B per year
Mobile Web advertising represents $36.7B per year
Desktop advertising represents $106.B per year
Applovin is effectively increasing their TAM by 40% by adding web-based advertising. This will take time to scale yet given the strong start we’ve seen on the in-apps ad business, to increase TAM by 40% is certainly something that catches our attention.
In terms of how this plays out, Applovin’s goal is to see 10% of revenue from the web-based business once the self-service platform goes live:
“After we launched the self-serve model, that business could grow quite significantly and outpace that 10% metric that we provided previously. So, it's quite likely that it could represent a larger than 10% portion of the revenue this year.”
Once the self-serve platform goes live, Applovin’s ambition is to see a market penetration in web-based advertising on par with their penetration in mobile games. The following was stated regarding the roll-out happening toward the back half of the year: “And then, even more exciting, I touched on this in the talk script, we're finally going to be releasing our new dashboard to some select advertisers for feedback this quarter. That's a huge catalyzing effect. When we do go to a full self-service state, we're going to open up our platform from a very small amount of advertiser penetration to the entirety of the world being able to come on to our platform.”
The company also clarified that until the self-service platform launches, they will not see meaningful revenue in the web-based business, yet in the meantime, there is “a line out the door.”
“We've got a line out the door of customers that have been waiting to come on to the platform, and then we've onboarded, I think I said, in a blog, hundreds of advertisers, but our team is small. So, when I say we're not looking to push, we're looking to push over time, but we need to get the self-service tools into the market so that we can pair that with the team to automate a lot more of the processes.”
Short Reports:
The CEO responded to the short seller reports here and here. These are worth a read as the CEO makes the explanation quite simple in terms of how their pixel is not unique and is aligned with industry standards.
Conclusion:
There are no guarantees in tech investing. Stick around long enough and you will see a bulletproof company fall flat (pick any high-flying best-of-breed cloud company – Zoom, Snowflake, Datadog, MongoDB), while others surprise the market for a decade or longer (Meta, Google, etc.).
Applovin gushes margins, has many catalysts to continue its growth trajectory and perhaps most importantly, is an underdog of sorts to where the market often (mistakenly)- affords a better entry in time. I won’t get any originality points for writing an analysis on Applovin after an 800% run in the markets in eighteen months, yet perhaps I can help by saying that from what I can see today, this run is not over yet.
The I/O Fund is closely monitoring Microsoft for a potential entry point. Join us Thursdays at 4:30 p.m. in our Advanced Market webinars, where we’ll outline our strategy for initiating a position with maximum upside in mind. Learn more here.Learn more hereLearn more here.
Pro Members: Don’t miss our biggest sale of the year — save $275 on an annual Advanced Market Signals plan. Email us to upgrade.Pro Members: Don’t miss our biggest sale of the year — save $275 on an annual Advanced Market Signals plan. Email us to upgradeEmail us to upgrade.
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.
TSS Inc. is an AI systems integrator partnered with a large U.S.-based OEM generating 99% of their 2024 revenue, presumably Dell Technologies for AI-enabled rack construction and integration.
TSSI’s revenue grew 523% to $99 million and adjusted EBTIDA grew 11X from $0.48 million to $5.2 million. Management's outlook indicates H1 2025 revenue will exceed H2 2024, with projected 50%+ year-over-year growth in Adjusted EBITDA for 2025.
TSSI stock surged from $0.24 to $14.49 in just over a year as it uplisted to the NASDAQ. The stock then cratered to as low as $6.24 during the tariff rout. Following its earnings report last week, the stock popped 70% — helping to illustrate the $300M market cap stock can be a wild ride.
TSSI more than doubled its headcount in 2024 and is doubling the size of its factory and headquarters as it moves into a 212,793-square-foot building in Q1 2025 to handle the extra capacity needed for its multi-year contract with its OEM partner (Dell Technologies) with guaranteed minimums.
There is speculation that TSSI is working on Elon Musk’s xAI Colossus supercomputer project through Dell Technologies, but the company won’t comment for the sake of confidentiality.
RISK: TSSI stock has a tiny 13 million share float and 25 million shares outstanding, which is at a high risk of material dilution as the company filed a $150 million shelf registration on January 7, 2025.
RISK: TSSI has no analyst coverage and 11.08% institutional ownership as of March 27, 2025,
RISK: Similar to crypto, TSSI requires technical analysis to be at the forefront of all buying and selling decisions. This stock is for advanced day traders who are comfortable with managing stocks daily due to high customer concentration and other notable risks.
The name TSS is an acronym for Total Site Solutions, which describes the nature of their business. The Texas-based distributor/reseller provides on-site installation, integration, deployment and confirmation services for data centers. Currently, the company is the partner of choice for Dell’s Integrated Rack Scalable Solutions business (IRSS).
TSS is situated in the sweet spot of the AI boom, helping customers build out their AI infrastructure. As Blackwell begins to ship in 2025, the complexity of the systems means that OEM companies like Dell and large AI companies like xAI will need assistance in integrating server racks, which includes sourcing components, assembling and integrating the racks, making sure the power needs are well balanced between liquid cooling and air cooling, and lastly, testing these systems and providing site surveys. All of this falls right into TSS's wheelhouse.
Company Background and Strategic Transformation (2019–Present)
Historically TSSI provided various data center solutions (integration, facilities management, and procurement of IT hardware), but they struggled to grow in any meaningful way. Gross profits increased from $7.21M in 2016 to $8.91M by 2022. The stock was around $.30 per share starting 2024 and had barely changed price since 2016.
In late 2022, TSSI began a major transition as Darryll Dewan became CEO and immediately focused on the high-value systems integration business. He was formerly a VP of global sales and marketing at Dell. Gross profits surged, with each quarter of 2024 seeing gross profit growth of 58.1%, 40.7%, 178.0%, and 122.9% YoY.
By the end of 2024 revenues reached $148M (up 172% YoY), gross profits reached $22M (up 100% YoY) and net income was $6M (versus essentially breakeven in 2023), all marking record highs for the company.
On November 14th 2024 TSSI announced a record 3Q and a multi-year agreement with their only customer, Dell. To support the multi-year agreement and Dell’s end market demand for AI servers, TSSI doubled its production footprint – relocating from its ~105,000 sq ft Round Rock facility to a new ~212,793 sq ft state-of-the-art integration center in Georgetown, TX (a 103% increase in space). Part of the agreement with Dell effectively guarantees that TSSI will at a minimum break even on the new expansion investment.
The new site offers a massive upgrade in power and cooling infrastructure to handle next-generation racks that will consume up to 6x more power than prior generations. TSSI invested ~$25–30 M in this build-out to future-proof its operations with liquid cooling test stations, heavy-duty flooring and lifts, and redundant power – all aimed at meeting the demands of AI racks at scale.
Management expects initial production in the new Georgetown facility by April 2025 and full production capacity by June 2025. The challenge and opportunity now will be scaling efficiently and executing to maintain their competitive position.
Due to the AI-fueled transformation TSSI underwent last year, the stock surged over 5000% from Jan 1st 2024 at $0.30 per share through its peak on Jan 23rd, 2025 at $15.22 per share before falling to $6.24 in April.
Background on Dell’s Rack Scale Solutions
Dell’s Integrated Rack Scale Solutions (IRSS) multi-year agreement essentially makes TSSI an extension of Dell’s manufacturing operations for AI servers. This is the primary driver to TSSI’s remarkable gross profit inflection in 2024. Thus, it is important to look more closely at TSSI's 99% customer.
Dell customers include tier 2 CSP’s like CoreWeave or Denvr Dataworks, the federal government, and large enterprises looking to build out on-prem data centers. Dell is not exposed to the traditional hyperscalers and so the growth forecast is not as correlated to hyperscale capex – although this could change and likely will given the rumors that Dell is working with xAI and also considering Super Micro may struggle to raise cash to quickly increase capacity (whereas Dell has operated at scale for decades).
Hyperscalers tend to do business for more customized AI server solutions rather than the turnkey solutions that Dell provides. Hyperscale vendors include companies like Taiwanese-based Quanta Computer or Wistron affiliate Wiwynn. However, it is interesting to consider the potential positive impact that tariffs may have on where hyperscalers source their AI servers going forward.
We still think that on-prem, tier-2 CSP’s and the federal government will drive a significant growth inflection for Dell’s AI server business in the near-term and durably grow for the foreseeable future. Dell’s calendar 2025 sales run rate is already trending to 50% YoY growth and 2x’d sequentially throughout FY2025. As such Dell’s AI server revenue could reach $20–30B by 2027, a ~40% CAGR from 2024 ($9B in AI server related sales).
While this multi-year agreement significantly reduces the uncertainty of TSSI’s future profits and revenue, the agreement can be terminated for convenience, meaning any party can opt out provided with 180 days written notice. Furthermore, if Dell terminates, Dell is no longer obligated to provide the minimum monthly volumes after the 180-day notice period, but they remain financially obligated to cover some of the costs associated with the Georgetown facility investment. In our view, the Dell agreement also comes with other limitations.
For one, it is unlikely that TSSI will have much negotiating leverage to increase prices. While they are investing in a new and upgraded facility, the reported $25-$30 million investment is a rounding error for companies like Dell or other large VAR’s/Distributors; hardly considered a barrier to entry given the cost and caliber of labor needed to assemble servers.
At the expiration of the multi-year agreement, there is a chance Dell does not renew or goes with another vendor. The investment provides TSSI some near-term advantages, yet the competitive positioning of TSSI is minimal.
SNX located in Taiwan is a significant competitor, for example, with $58.5 billion in annual revenue. The Hyve business has a large, global footprint that competes with TSSI. Perhaps Dell initiated an investment in TSSI to further localize systems integrations and procurement well ahead of tariffs (time will tell).
Dell to See Calendar Year H2 2025 Ramp:
The IOF Fund article, “Dell Q4: Projects $15 billion in AI shipments this year”, Dell Q4: Projects $15 billion in AI shipments this year”, noted that Dell’s FQ1 2026 (ending May 2, 2025) revenue guidance missed consensus analyst estimates by 3% and EPS guidance of $1.65 missed analyst estimates for $1.78. In its FQ4 2025 (ending January 31, 2025 calendar year), Dell shipped $2.1 billion in AI servers (down 28% QoQ), with orders at $1.7 billion (down 53% QoQ) and a backlog of $4.1 billion.
Management guided AI shipments of $15 billion in FY F2026 as AI server backlog doubled to $9 billion in FQ4, primarily driven by recent deals, including xAI. Management’s $15 billion FY F2026 shipments guidance implies the NVL delays will make it a second-half story for AI ramp-up:
“Where Dell and Super Micro may both be seeing lower growth than expected likely goes back to the delivery of key Nvidia systems, where the larger systems lead to higher revenue (and you’re aware by now these were delayed by “couple months”). It’s also perhaps due to Nvidia’s partnership with Foxconn, who has seen more news lately than peers Dell and Supermicro in terms of shipping Blackwell systems. According to a news report from Economic Daily the GB200 was shipped by Foxconn in small quantities at the end of Dec and is expected to be shipped in large quantities at the end of January.”were delayed by “couple months”). It’s also perhaps due to Nvidia’s partnership with Foxconn, who has seen more news lately than peers Dell and Supermicro in terms of shipping Blackwell systems. According to a news report from Economic Daily the GB200 was shipped by Foxconn in small quantities at the end of Dec and is expected to be shipped in large quantities at the end of January.”
Across the board, key Nvidia suppliers like Dell should see a strong ramp into the second half of the year, with this flowing down to TSSI, especially as projects ramp in size (such as xAI’s Colossus).
Elon Musk’s xAI Supercomputer Project
TSS CEO Darryll Dewan, formerly VP of Global Sales and Field Marketing at Dell from 2012 to 2022, commented in the Q2 2024 conference call, “Demand increased in Q2, and we began delivering complex AI integration solutions on time, and I want to stress, on time, including the first stage of a highly publicized program. That initial program began in June and is being carried out into Q3 [2024]. As a result, we finished a quarter with a record run rate of rack integration revenue.” Dewan stated that the volume ramp they had been anticipating was underway, and its Q2 results were a harbinger of things to come. Dewan also would not confirm during the Q&A session when asked directly if xAI was one of their projects.
The “highly publicized program” is speculated to be Elon Musk’s xAI supercomputer “Colossus” project. The xAI data center houses 100,000 GPUs comprised of over 1,500 racks and received approval for 150MW of power, enabling all GPUs to run concurrently.
Dell Technologies is Involved in Assembling Half of xAI’s Racks
Musk already revealed in June that Dell Technologies is assembling half of the racks going into the supercomputer project and Super Micro Computer would also be involved. It's been speculated Elon Musk shifted $6 billion in AI server orders for xAI to Dell Technologies and away from Super Micro Computer due to their accounting issues. This trickles down to TSS.
Musk had also announced the expansion of Colossus to 200,000 GPUs in October, and there is growing speculation at the moment that xAI is currently exploring a fundraise of tens of billions of dollars for the buildout of “Colossus 2’, which is rumored to include as many as 1 million GPUs, or 10x the size of the original Colossus supercomputer. In February, it was rumored that Dell had won a $5 billion AI server deal with xAI for Nvidia’s GB200 platform, though it is unclear whether this is for Colossus or Colossus 2.
TSS’s Q3 2024 is assumed to have contained a whole quarter’s worth of xAI business, which could be a sign of things to come. As Dell’s AI Factory server business ramps up, so does TSS’s business, as evidenced by Dewan’s statement, “Volume expectations are dependent on sales execution by our OEM partner, but our partner has shown great confidence in TSS by committing to help to smooth what otherwise could be a feast or famine business.”
Upgraded Headquarters to New Site:
Similar to other AI-driven companies, TSS insists it does not have demand issues. Rather, it's a question of how quickly capacity can be added, with TSS upgrading its facilities for AI rack integration services. TSSI expects the new facility to reach full production capacity in June 2025.
On January 5, 2025, TSS announced it signed a long-term lease for a larger facility with 212,793 square feet, essentially doubling its earlier space, which was 105,000 square feet. TSS is moving its headquarters to the new factory located in Georgetown Logistics Park in Georgetown, Texas, which will be online in Q1 2025. In the six months leading up to its August 14, 2024, conference call, CEO Dewan confirmed they have more than doubled their headcount.
The company has stated site plans call for a $25 million to $30 million investment for improvements to bring additional power to the building which will provide greatly expanded cooling capacity for its rack testing and validation stations. It will triple the capacity to test and validate direct liquid-cooled racks in addition to traditional air-cooled racks. There is no doubt the industry is migrating to liquid-cooled rack technology.
CEO Dewan commented, "Continuing our rapid growth trajectory was centered around two key drivers: signing a long-term agreement with our primary customer, which we completed and announced in October, and building capacity to deliver the demand driven by AI infrastructure needs in the market. Our new facility more than doubles our square footage and positions TSS to continue our rapid growth. We are beginning the required fit-out immediately and expect to be operational in the new building in the first quarter of 2025." Management updated in Q1 that the buildout was progressing according to its plan, and that the built-out capacity was higher than current demand, allowing them to scale higher in the future.
When asked if capacity is a limiting factor, the management team stated they “…have the capacity to grow 10X” and “so, the timeline, a couple of years may be before we start to get a little tight.” That is music to an investor’s ears, yet power requirements remain a constraint (and perhaps a tailwind for TSSI).
The Role TSSI Plays in Increased Power Consumption of AI Data Centers
The new facility building was originally planned for 4.5 MW but will now begin with 6 MW and 15 MW by early summer, and 40 MW over time. The 15 MW timeline for early summer was reiterated in Q1, with management noting this would be ~6x their current facility in Red Rock. Just as with upgrading the power, the cooling situation also had to be upgraded. CEO Daryll Dewan said this in their Q4 2024 conference call.
“Cooling is in a similar situation. When we began the fit-out of our facility, it was anticipated we would integrate a mix of chilled air and direct liquid-cooled technology. However, the adoption of emerging chip families so quickly has resulted in an accelerated shift to direct liquid-cooled. This impacts everything from our chiller capacity to the diameter of the pipes coming into the facility and distributing water within the facility. And again, this rethinking has all occurred in weeks.”
AI is causing an unprecedented surge in power density at data centers with current AI racks pushing 80 kW, moving to 120 to 150 kW, and eventually 200 kW in the next few years. The I/O Fund has covered the generational leap in power consumption in our blog article, ”AI Power Consumption: Rapidly Becoming Mission-CriticalAI Power Consumption: Rapidly Becoming Mission-Critical.”
Rapid increases in power requirements not only create potential failures and raise costs but there is also the challenge of increasing compute density in data centers. TSS is positioned to help customers make nimble, on-the-fly architecture changes, including cooling options, thanks to its rapid testing capabilities, which can narrow configuration options. CEO Dewan stated in its Q2 2024 conference call, “But importantly, the next generation of racks will consume up to 6x more power than those being produced today.”
Note on Modular Data Centers:
Modular data center revenue grew 13% in 2024. In addition to integrating a combination of air-cooled and liquid-cooled racks, TSS has also configured and deployed over 350 modular data centers (MDCs), which are pre-fabricated and scalable portable data centers. Due to the long lead times to build and deliver specialized data centers, the demand for MDCs is expected to grow as AI adoption grows. These carry gross margins north of 50% and grew 13% YoY in 2024, as it’s a predictable revenue stream.
Here is what was stated in the earnings call:
“We're also very excited about some of the conversations we're having about different design points on the modular unit. I'd like to go into a little bit more detail, but I don't think it'd be appropriate. But I think where we can provide an AI solution to an enterprise to deploy a certain amount of power cheaper, better, faster than their alternative. And that alternative could be a co-lo, could be a hyperscaler, could be expanding their own existing data center space.”
Financials Overview: Revenues Accelerate in Q1
Q1 revenue nearly broke into the triple-digits as it accelerated significantly from Q4’s 105% YoY growth to 523% YoY growth. This was driven primarily by Procurement revenue, which rose nearly 7x YoY, while management stated that there was incremental contributions from AI rack integration services. This is only the third full quarter of contribution from AI rack integration services after commencing this in June 2024.
While TSSI did not provide a guide for Q2, management stated that they expected 1H 2025 revenue to outpace 2H 2024, where they generated just over $120 million in revenue. As it stands, Q2 would need to have just $22 million in revenue to meet that forecast, though underlying business momentum suggests that is far too low.
Key Segments
Here’s how TSSI’s revenue by segment looks, with the AI rack-driven System Integration segment beginning to perk up though Procurement revenues drive the bulk of TSSI’s revenue.
Procurement Services Revenue Surges 676%
This segment consists of sourcing and selling third-party hardware, software, and services to customers – effectively acting as a value-added reseller (VAR) or distributor. It is TSSI’s largest segment by revenue but lowest by margin. It is a very lumpy business due to seasonal spending trends by the federal government being 2H weighted, and not expected to be a major profitability driver for TSSI due to its thin margin profile. Management sees Procurement as a strategically important complement to Systems Integrations, as deals “often bundle or precede integration projects.”
Procurement is uniquely impacted by how deals are recognized, either as gross or net: a gross deal occurs when TSS takes ownership of the hardware as they transform the product and record the gross value of the transactions as well as the gross cost of the goods, resulting in higher gross revenue but lower gross margins between 3% to 4%. A net deal is when TSS acts only as an agent in buying and selling the product, as they only record the agency fee, resulting in a 100% margin. TSSI added that there was also a higher mix of gross deals in the quarter, which weighed on margins.
Procurement revenue accounted for more than 91% of revenue in Q1, as it surged 676% YoY to $90.2 million, aligning with management’s Q4 projection that revenue would remain elevated for 1 to 3 quarters. This was driven by increased purchases by the federal government, as well as a few individually large sales to commercial enterprises in the quarter to support AI workloads. Additionally, some, not all, of Procurement revenue flows through to Systems Integration as it relates to components needed for AI and non-AI server racks.
Systems Integration Revenue Ramping
Systems Integration (SI): This is TSSI’s flagship segment and growth engine, encompassing the design, assembly, and testing of integrated technology solutions – most notably high-performance computing racks for AI and other advanced workloads. This business involves taking servers, GPUs, networking gear, power/cooling components, and software, and building turnkey rack systems to customer specifications. It is a project-based, engineering-intensive service and carries higher margins. This segment functions similarly to a company like Super Micro in that TSS assembles servers and server racks as a vendor.
The higher-margin, AI rack focused System Integration segment is beginning to see growth ramp in the third quarter of AI rack integrations, operating currently at a ~$30 million annual run rate. This builds on strong growth from 2024, which saw segment gross profit rise 480% YoY on a 157% YoY increase in revenue.
Q1 revenue increased 252% YoY to $7.5 million, its third consecutive quarter of >200% YoY growth driven by AI rack integrations. TSSI says that it receives both fixed monthly fees as agreed upon in the multi-year agreement, as well as “payments that scale depending on the volume of AI racks integrated and for which we are prepared to integrate.” TSSI’s order pipeline from Dell remains “extremely robust.”
For the non-AI rack integration side, TSSI noted in the 10-Q that it may be impacted by supply chain issues or lulls in demand, impacting revenue as it waits for delivery of certain required components. TSSI said that its vendors and partners expect “supply-chain issues to continue for at least the next several quarters, though they appear to be improving in general.”
Management said in Q1’s call that they “expect sustained high growth in this area as customers ramp-up investments to meet evolving compute demands over the coming quarters and years,” with this segment expected to become the primary growth driver.
According to an internal model, by the end of 2027, SI could be roughly 63% of total gross profits, up from 15% of total gross profits in 2023. As a precaution, we’ve modeled 62M in revenue and 30M in gross profits in 2027 for the SI segment. This is at the low end of the 2-4x capacity comment and assuming a 1:1 relationship between volume and sales, but this forecast could have significant variability due to high customer concentration.
In this internal forecast, there is a degree of conservatism given the project-based nature of this business and the range management provided of 2-4x 2024 peak volume capacity. Dell is guaranteeing at least as much volume as TSSI’s peak quarterly run-rate in 2024 for the SI segment. For context, TSSI’s peak in 2024 was around Q4 when SI revenue hit $7.9M. Assuming that there is a 1:1 relationship between volumes and revenues, this could translate to a baseline of $31M in annual systems integration sales and $13M gross profits which is based on the 2H2024 run-rate.
As such, the incremental sales potential for SI is $62M–$124M in sales and $26M–$52M gross profits. Again, this is with what we know today, yet carries significant variability due to reliance on one OEM.
Facility Management Revenue Declines
This segment involves data center facilities services – including maintenance contracts, on-site support, and deploying modular data centers. It provides a steady, recurring revenue stream, and while it has high gross margins, it is not going to be the real driver of TSSI’s profitability inflection, which squarely falls onto the SI segment. It has grown in the 10% revenue range over the last several years. Modular data centers are important and they will grow in volume, but this segment is more of an indirect beneficiary of the AI boom. Furthermore. TSSI is not constructing these data centers, but rather maintaining them, fixing them, and monitoring them.
Facility management revenue declined (40%) YoY to $1.3 million, with management noting they are currently optimizing this unit to focus on high-growth opportunities. Facility management had grown just 13% YoY in FY24, a far cry from System Integrations’ 157% growth and Procurement’s 205% growth.
Management added that they “anticipate more robust growth over the next 12 to 18 months” as medium and large enterprise customers “increasingly adopt modular data centers as a cost-effective solution to leverage AI technologies.” For 2025, we estimate $8.7 M (+9% YoY), assuming TSSI continues to win small expansions or projects. Management mentioned a couple of projects drove 46% growth in Q2 2024 facilities revenue, but on a normalized basis, mid-single digits is more in-line with their historical multi-year rate of growth.
Margins Weighed Down by Procurement Growth
TSSI has noted previously that margins are likely to fluctuate quarter-to-quarter as revenue mix shifts, and Q1’s heavy concentration of Procurement revenue with higher mix of gross deals weighed on gross margin, though operating margin felt less of an impact. This is because a portion of Procurement revenue flows through to the Integration segment, which carries far higher margins.
Q1’s gross margin was 9.3% as a result of this mix shift, down more than 5 points sequentially and nearly 8 points YoY.
Procurement GAAP gross margin was 7.8%, flat YoY; on a non-GAAP basis, which strips out gross vs net deals, gross margin was 6.6%, up 2 points YoY.
Systems Integration gross margin was 22%, down 6 points YoY due to a $0.8 million rent expense for the new Georgetown facility; stripping out this noncash rent impact, gross margin was 32%. Management said that they expect gross margins in the segment to improve in the last three quarters in FY25.
Facility management gross margin was 41%, down 15 points YoY on lower revenue.
Q1’s operating margin was 4.2%, down only 1.3 points QoQ and up 2.6 points YoY. Should Systems Integration begin to grow its share of revenue with margins growing, TSSI should theoretically see some gross and operating margin expansion as the year progresses.
GAAP EPS Expanding
Q1’s GAAP earnings were $0.12 per share, surpassing Q3’s earnings of $0.10 despite a thinner margin profile this quarter. This was a notable improvement from the break-even quarter last Q1.
There’s also a notable inflection in earnings power since the start of AI rack integrations in June 2024, with EPS over the past three quarters of $0.30, up 10x from $0.03 in the comparable period last year.
Cash and Balance Sheet
Cash flows have been quite variable, with Q3 2024 and Q1 2025 seeing large positive operating cash flow while Q4 saw a large outflow.
Operating cash flow was $20.6 million in Q1, up nearly 8x YoY and a stark contrast to the ($21.6) million outflow in Q4 2024. OCF margin was 20.9%, up more than 4 points from 16.7% a year ago.
Free cash flow was $5.8 million, up ~2.2x YoY and again a stark contrast to the ($28.4) million from Q4. FCF margin was 5.9%, down from 16.4% a year ago due to the rapid revenue growth.
Adjusted EBITDA was $5.2 million for a 5.3% margin in Q1, up from $0.5 million for a 3% margin in the year ago quarter. Management reiterated its view for >50% growth in adjusted EBITDA for 2025, signaling they expect at least $15.3 million this year.
Cash and equivalents totaled $27.3 million, while debt was $8.2 million. Debt is likely to be higher next quarter as TSSI said it drew down the remaining $11.3 million on its construction loan just prior to Q1’s call.
Factoring Receivables is an Expensive Payday Loan Approach to Stabilize Cash Flow
TSSI sells Dell’s invoices/receivables to a third-party factoring company at a slight discount and interest in exchange for immediate cash, similar to a payday loan. This is a common strategy to improve cash flow with large clients, the largest client in this case, with longer payment terms. It’s an expensive way to get paid sooner. TSSI sells Dell’s invoices to the factoring company at a small discount, and the factoring company immediately wires over cash to TSSI. Interest is applied to the wired funds until Dell pays off the invoice with the factoring company, which then wires the remaining balance back to TSSI.
In Q3, the net interest expense was exclusively the cost of factoring Dell’s accounts receivables, which has an effective interest rate of 6%, a lower rate than a bank loan. In Q4, factoring interest expense was $721,000, before more than doubling QoQ to $1.5 million in Q1. This continues to illustrate how vital Dell is to TSS’s future and the leverage it holds.
Conclusion
TSSI’s tie-ins with Dell and ramping AI server rack integrations as more production capacity comes online support strong revenue growth, and the company remains profitable despite high Procurement mix weighing on margins. While AI servers have the potential to drive meaningful growth for TSSI, the small float and market cap lead to increased volatility, requiring active management.
The I/O Fund has a high allocation to other prominent data center infrastructure beneficiaries, sharing its research and real-time trade alerts with our Pro and Advanced subscribers, while discussing potential setups and trading plans in our weekly webinars with Portfolio Manager Knox Ridley. 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. Take advantage of a limited-time offer for $275 off our Advanced tier herehere. To upgrade, email us at premium@io-fund.compremium@io-fund.com.
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.
