Super Micro beat Q4 revenue only marginally by $10 million, yet the Q1 and full year fiscal 2025 revenue outlook came in far ahead of estimates. Fiscal Q1 guide beat consensus by more than $1 billion with management guiding between $6 billion and $7 billion for next quarter, compared to consensus of $5.45 billion. This will represent record revenue growth of 207% at the midpoint. The previous highest growth rate was 200.7% two quarters back. We’ve called Supermicro the AI bullet train, and it’s quite clear that train is still in motion.

Management also guided on the call for fiscal year 2025 revenue between $26 billion and $30 billion, which is 30% higher than consensus. Going into this call, FY2025 revenue growth was expected to be 57.2% and will now be 87.4%, at the midpoint.
Looking beyond the impressive acceleration in revenue growth for Q1 and the welcomed raise for next year, Supermicro on a company-specific level has weak margins and a cash flow problem. The margins also got hit hard due to the cost of direct liquid cooling components.
We’ve been quite clear the cash flow is this company’s Achilles heel, and this was the key reason we closed the position in the past. However, the valuation is low enough (and the growth is high enough) that odds favor the stock bottoming soon.
Bullish Liquid Cooling Growth
Here’s a quick refresh on liquid cooling from the recent analysis we published on Liquid Cooling Leaders:
“Liquid cooling technology has been around for decades, yet this technology is becoming mission critical due to the increasing levels of compute power from AI accelerators, starting with the GB200 systems and B200 GPUs […] Although the GB200 will ship end of this year, and the B200 will fully ship in early 2025, vendors are scaling their liquid cooling capacity now […]
The Blackwell architecture is a catalyst for liquid cooling as it nears 1000W, specifically the GB200 systems and the B200. This represents a 40% increase from the previous generation.”
At the time, we discussed Blackwell being a clear catalyst for direct liquid cooling. Therefore, it is odd to say the least that Supermicro would report that DLC is surging when there are rumors that Blackwell is delayed. Here are snippets as to what the CEO discussed on Supermicro’s call in terms of how quickly DLC is ramping:
“Answer
Charles Liang (Executives):
Yes. Thank you. I mean as you know, liquid cooling has been in the market for 30 years and market share compared with overall data center size always small, less than 1% or close to 1%, I would have to say. But just June and July 2 months alone, we shipped more than 1,000 racks to the market. And if you calculate 1,000 racks, AI rack right, is about more than 15% on a global data center new deployment”
To another question, the CEO responded:
“It's a very good question. I mean, last month, we have about 1,000 racks per month liquid cooling capacity. And today, we already grow another 50%. So now we have a 1,500 rack per month capacity. By this year-end, we will grow that to 3,000 rack per month.” Takeaway: that’s 50% growth in one month and 200% growth in 6 months on liquid cooling.
Regarding Blackwell, the CEO stated:
“So for Q3, for sure, we do not expect any Blackwell volume. For Q4, I mean December quarter, I guess, it will be very small. Engineering sample small volume. So the real volume, I believe, had to be March quarter next year. And that's why we only $26 billion to $30 billion.”
The CFO also added: “I would say, we prepared the market for a downturn in margins or a softening of margins in our guidance last quarter. But even we were surprised by the acceleration that we saw in the liquid-cooled rack market. And so we had to ramp up our supply chain. We paid a lot of expedite costs and higher supply chain costs. So I think as the supply chain improves, we expect those efficiencies to now come back out, but that impacted us more than we had expected.”
Putting the Pieces Together:
Supermicro has to walk a fine line and cannot speak for Nvidia. However, per Tom’s Hardware and other sources, direct liquid cooling truly starts with Blackwell: “Even Nvidia's high-end H100 and H200 graphics cards work well enough under air cooling, so the impetus to switch to liquid hasn't been that great. However, as Nvidia's upcoming Blackwell GPUs are said by Dell to consume up to 1,000 watts, liquid cooling may be required.”
Supermicro’s report is communicating that servers that require direct liquid cooling are soaring (suddenly) as of June and July from 1% of all new servers shipped to 15% at 1,000 racks. Management is also communicating that it’s expected to continue to soar to 3,000 racks by the end of this year. Per our understanding and industry analysts (like Tom’s Hardware) there’s an incredibly high probability this is Blackwell driving a sudden surge in DLC sales. But if Blackwell is delayed, how can this be?
We will need Nvidia’s report to be certain, but one possibility is the GB200 NVL36 and NVL72 systems are taking up TSMC’s CoWoS-L capacity. That implies there would still be Blackwell sales through the GB200NVL systems, while pushing the B200 release out further. Essentially, one possible theory I’ve come up with is that the larger systems with 36 GPUs or 72 GPUs are so popular, the single B200 GPUs are being delayed.
This theory is supported by preliminary data that the GB200 systems were reportedly seeing outsized demand, which we’ve shared on social media here, here and here. Theoretically, if the GB200 systems were seeing outsized demand (per the preliminary data), it would bump the B200s to a later date.
Here's a breakdown on how pricey the GB200 systems are:
- Nvidia’s GB200, featuring one Grace CPU and 2 B200 GPUs, is estimated to sell for ~$60,000 to $70,000.
- In the NVL36 configuration, featuring 18 GB200s (18 Grace CPUs and 36 B200s), each GB200 would be selling for $100,000 at the current estimated ASP of $1.8 million.
- In the NVL72 configuration, featuring 36 GB200s (36 Grace CPUs and 72 B200s), each GB200 would be selling for ~$83,333 at the current estimated ASP of $3 million.
In this case, Nvidia would theoretically prioritize the GB200 NVL36 and NVL72 as the price points are quite high. Per Semi Analysis: “Combine these two issues and it’s clear that TSMC will not be able to supply enough Blackwell chips as Nvidia would like. Consequently, Nvidia is focusing what capacity they have almost entirely on GB200 NVL 36×2 and NVL72 rack scale systems. HGX form-factors with the B100 and B200 are effectively now being cancelled outside of some initial lower volumes.”Nvidia is focusing what capacity they have almost entirely on GB200 NVL 36×2 and NVL72 rack scale systems. HGX form-factors with the B100 and B200 are effectively now being cancelled outside of some initial lower volumes.”
The two NVL36 and NVL72 rack configurations carry a ~27% to ~54% higher selling price per GB200, making it understandable why Nvidia would focus on the racks given production constraints from CoWoS capacity.
This is one theory as to how Supermicro could forecast a surge in DLC server shipments, and meanwhile, the B200 GPUs be delayed.
Another explanation is that the rising costs of power consumption is causing customers to order DLC servers instead of air cooled for the H100 and H200. Yet, it’s the sudden surge in sales (in two months) that has me leaning toward the first possibility where SMCI’s 1,000 racks is from the GB200 NVL systems that can be produced and are still shipping, with GPU clusters are being prioritized over GPUs.
Going into the report, we had stated: “What we hopewe hope to hear from management is that the exact date for the B100s and B200s arrival is immaterial given the demand environment. Meaning, there are enough buyers lined up for Nvidia’s Hopper GPUs, along with enough CoWoS-L capacity from TSMC to meet demand for the GB200’s, that combined this can meet or exceed fiscal year estimates. There are also additional variations of Blackwell being designed for the CoWoS-S packaging from TSMC.”
Supermicro confirmed today the company is able to meet and exceed current estimates regardless of the current Blackwell delay.
Why the Stock Sold Off
Despite the incredibly bullish commentary on direct liquid cooling, the stock reversed from being up double-digits to being down double-digits. While the headline numbers point toward it being the margins, my hunch is it’s the cash situation.
Here are two questions from the call that echoes our concerns a few months back:
Question
Dong Wang (Analysts)
Okay. Can you address on the working capital, if you can give any color on that?
Answer
David Weigand (Executives)
Yes. So we announced a $500 million credit line with a group led by the Bank of America. And so we expect we are really working on our balance sheet and leveraging our balance sheet. And we expect to — some announcements to be coming in terms of additional loan possibilities in the future.
Question
Mehdi Hosseini (Analysts)
And then a question I have for Charles. Obviously, you've done a good job of doubling revenue in fiscal year '24. But you also had a negative free cash flow of $2.6 billion. And if I were to look at the high end of your revenue guide for fiscal year '25, you're on track to double revenues again. Does that mean that you're going to need to burn another $2.5 billion to $2.6 billion of free cash flow to hit those revenue targets?
Answer
Charles Liang (Executives)
Not necessarily. I mean, if we try to be very aggressively growing market share, maybe — for example, we forecast $30-something billion, right, so in that case, we may need more. But if we try to focus on below $30 billion, then not necessary.
Answer
David Weigand (Executives)
And Mehdi, 1 thing I would add to that is we believe that we have NIG profile, and as such, like I mentioned earlier, we're starting to leverage our balance sheet more with targeting toward unsecured debt. And so that will help us on an inter-quarter basis.”
My takeaway: Cash remains Supermicro’s Achilles heel and per our last analysis: “What lies beneath this phenomenal growth rate is the need to raise cash to fund operations, which for Super Micro means buying excess inventory to prepare for future growth, especially as it relates to liquid cooling.”
However, when I wrote that in April, the company was trading at a 3.3X Forward PS with a Forward PE ratio of 36X. With the new fiscal year guide, we are in the 1.3X forward PS range and 15 forward PE range. There is a lot of negativity priced into Supermicro right now, and when you account for the cash situation, I think we are close to a bottom of sorts. The economy is out of an investor’s control, yet identifying quality companies at a discount is one way to combat the volatility.
Revenue and EPS:
As discussed, Q4 marked a third consecutive quarter of triple-digit top line growth, and Q1’s guide points to a more than 60 percentage point acceleration to >200% revenue growth. Next quarter will also mark the highest quarter growth in Supermicro’s history. This quarter growth was driven by: “strong demand for next-generation air-cooled and direct liquid cooled rack scale AI GPU platforms, representing over 70% of revenues across enterprise and cloud service provider markets where demand remains strong.”

- Q4 revenue was $5.31 billion, an increase of 144% YoY and 38% QoQ. This compares to consensus for $5.30 billion in the quarter.
- Management guided for a strong Q1, projecting revenue between $6.0 billion and $7.0 billion, for a YoY increase of nearly 207% at midpoint and well ahead of the consensus estimate for $5.45 billion.
- FY24 revenue was $14.94 billion, an increase of 110% YoY.
- For the full year, Super Micro guided for revenue between $26 billion and $30 billion, for YoY growth of 87%. This was nearly $5 billion ahead of the consensus estimate of $23.4 billion at midpoint, suggesting that Q1’s revenue level is expected to be the floor for the year.
Earnings:
- GAAP EPS was $5.51 in Q4, an increase of nearly 61% YoY but a QoQ decline of (16%) as gross margins shrunk sequentially. Per management: “Some key new component shortage delayed about $800 million of revenue shipped into July, which lowered our EPS for June and will be recognized in our September quarter.”
- Adjusted EPS of $6.25 missed estimates for $8.14, primarily due to margin weakness in the quarter. Adjusted EPS increased more than 78% YoY but declined (6%) QoQ.

- Q1’s GAAP EPS was guided at $5.97 to $7.66, or ~$6.82 at midpoint for YoY growth of 148%. Adjusted EPS was guided at $6.69 to $8.27, or ~$7.48 at midpoint for YoY growth of 118%; this was slightly below estimates for $7.68 in the quarter.
- GAAP EPS was $20.09, an increase of nearly 76% YoY. Adjusted EPS was $22.09, an increase of 87% YoY.
Margins:
This is where Q4’s report struggled, with gross margin contracting to a low 11% level, driving the bottom line miss. Regardless, of the uptick expected next quarter to 12%, this is a concern as DLC components drag on the margins. Operating margins also contracted significantly in Q4.
- GAAP gross margin was 11.2% in Q4, reaching the lowest level ever, and contracting from 15.5% last quarter and 17% in the year ago quarter. According to management, gross margins is expected: “to be above 12% in the first quarter.”
- Adjusted gross margin was 11.3%, down from 15.6% last quarter and 17.1% in the year ago quarter. This adjusted gross margin level was in line with the prior two quarters adjusted operating margin, highlighting the weakness in gross margins this quarter.
- GAAP operating margin was 6.5%, down from 9.8% last quarter and 10.4% in the year ago quarter.
- Adjusted operating margin was 7.8%, down from 11.3% last quarter and 11% in the year ago quarter. Per the CFO: “which is lower than what we expected due to the higher mix of hyperscale data center business and expedited cost of our DLC liquid cooling components in June and September quarter.”
- GAAP net margin was 6.7%, down from 10.5% last quarter and 8.9% in the year ago quarter.
- Adjusted net margin was 7.6%, down from 10.7% last quarter and 9.2% in the year ago quarter.
For the full year, margins shrank rather significantly:
- FY24’s GAAP gross margin was 14.0%, down from 18.1% in FY23. Adjusted gross margin was 14.1%, down from 18.2% in FY23.
- GAAP operating margin was 8.5%, down from 10.7% in FY23. Adjusted operating margin was 10.0%, down from 11.4% in FY23.
- GAAP net margin was 8.1%, down from 9.0% in FY23. Adjusted net margin was 9.0%, down from 9.5% last year.
Cash and Debt:
Super Micro recorded another quarter with significant cash outflow, and for the full year, operating cash flow was roughly ($2.5 billion), driven by increasing inventories, which rose by $3 billion YoY to $4.4 billion by the end of Q4.
- Operating cash flow was ($635 million), for a margin of (12%); this was a notable improvement from Q3’s outflow of more than ($1.5 billion), but still marked a third straight quarter with significant cash outflows.
- For FY24, operating cash flow was ($2.48 billion), a substantial change to OCF of $664 million in FY23. This represented an annual OCF margin of (16.6%).
- Free cash flow was ($662 million), with capex of just $27 million, far below management’s expectations for $55 to $60 million in capex in the quarter.
- FY24 free cash flow was approximately ($2.61 billion), for a margin of (17.5%).
- Cash and equivalents totaled $1.67 billion at the end of Q4, despite Super Micro padding the balance sheet to $2.12 billion in cash at the end of Q3. Burning through this much cash this rapidly raises the risk of another capital raise.
- Inventories were $4.4 billion, up from $1.4 billion at the end of FY23 and
- Debt and convertible securities totaled $2.17 billion.
Conclusion:
If you track Supermicro’s commentary closely, the direct liquid cooling market has accelerated from previous expectations only two months ago. At Computex 2024 in early June, CEO Liang stated he “expects 15 percent of racks it sells this year to use DLC, and 30 percent to employ it next year.” Fast-forward only two months and the company is now stating: “And if you calculate 1,000 racks, AI rack, it's about more than 15% on a global data center new deployment” [shipping now] and “we are targeting 25% to 30% of the new global data center deployments to use DLC solutions in the next 12 months.”
It's a mystery as to how Supermicro could achieve this if Blackwell is truly delayed the way the media portrays it (the media is ultimately portraying that the delay means a loss of revenue). We’ve provided one theory, which is that the high priced GB200 systems’ popularity has crowded out the other SKUs. We won’t know for sure until Nvidia reports if the GB200 NVL systems are shipping as planned, but we do know for sure that Supermicro is accelerating in revenue next quarter and raised guidance by 30 points for next fiscal year. It is the mark of a multi-generational opportunity that you could have any sort of delay, and yet demand is so great, there is still an acceleration in growth from the server maker.
As previously stated, there is a lot of negativity priced into Supermicro right now, and even when you account for the company needing to raise cash to support growth, I think we are close to a bottom of sorts for this stock. The economy is out of an investor’s control, yet identifying quality companies at a discount is one way to combat the volatility.
Damien Robbins, Equity Analyst for the I/O Fund, contributed to this analysis.
Recommended Reading:
- Super Micro FQ4 Preview: High Anticipation for Blackwell & DLC Commentary
- Cloudflare Q2: Significant Margin Expansion, Customer Acceleration
- Microsoft Fiscal Q4 2024 Earnings: Capex Surges QoQ; Azure Remains Durable
- AMD Q2: Data Center Accelerates to Growth of 115%
- Liquid Cooling Leaders: Super Micro, Dell, Vertiv and HPE