AMD Q2 Earnings: Key Updates About AI

Summary:

• Advanced Micro Devices, Inc. reported Q2 earnings and provided updates about its data center GPU business.
• AMD’s performance in this segment remains strong so far this year, and the company should see solid growth through the rest of the year.
• However, I remain concerned about AMD’s ability to sustain its growth trajectory in 2025 due to intense competition from Nvidia.
• Overall, I maintain a Hold rating for AMD stock.

Advanced Micro Devices, Inc. (NASDAQ:AMD) reported Q2 earnings yesterday (July 30). The company beat on revenue and was roughly in line with expectations on EPS. On the earnings call, management also provided updates about AMD’s data center GPU business, including how its MI300 chips are doing, as well as some information about the company’s upcoming GPU roadmap. In this article, I take a closer look at the most significant updates from AMD and consider their implications for the company’s future prospects in the data center GPU market.

Q2 MI300 Revenue Exceeded $1 Billion

For the last few quarters, I have been projecting AMD’s quarterly data center GPU revenue based on the company’s updates about its MI300 backlog along with commentary from management. I last updated these projections after Q1, arriving at the following revenue figures for 2024 (for a total of $4.3 billion):

• Q1: $600 million (actual)

• Q2: $900 million

• Q3: $1,200 million

• Q4: $1,600 million.

In Q2, AMD came in ahead of my projection by over $100 million. Here is what management stated on the earnings call:

Turning to our data center AI business, we delivered our third straight quarter of record data center GPU revenue with MI300 quarterly revenue exceeding $1 billion for the first time.

This is a solid beat, representing outperformance of more than 11% compared to my earlier projection. AMD’s overall revenue beat by $120 million in Q2, so it seems to me that almost the entirety of the company’s revenue beat this quarter may have been driven by outperformance in the data center GPU business.

Increased MI300 Backlog For 2024

AMD has been providing backlog figures for the MI300 since late last year. The figure stood at $2 billion in October, $3.5 billion in January, and $4 billion in May. This quarter, the figure was increased by another $500 million or so. Here is what management had to say on the earnings call:

Customer response to our multi-year Instinct and ROCm roadmaps is overwhelmingly positive and we’re very pleased with the momentum we are building. As a result, we now expect data center GPU revenue to exceed $4.5 billion in 2024, up from the $4 billion we guided in April.

Management also expects that MI300 revenue will continue to ramp in both Q3 and Q4. Based on these pieces of information, I am now updating my 2024 quarterly MI300 revenue projections to the following figures (for a total of $4.6 billion):

• Q1: $600 million (actual)
• Q2: $1,000 million (actual)
• Q3: $1,300 million
• Q4: $1,700 million.

Based on these projections, AMD should reach a run-rate of about $6.8 billion/year exiting 2024, which is certainly quite a lot of progress considering that AMD’s data center GPU revenues in previous years had been quite negligible.

Moreover, it appears that the growth in data center GPU sales over the last year has generated a pretty significant boost for AMD’s gross margins (and through them, profitability). As management noted:

Gross margin was 53% [in Q2], up 340 basis points year-over-year, primarily driven by higher data center revenue.

As MI300 revenue grows further over the remainder of the year, it seems reasonable that gross margins and profitability will likely see a further increase.

The Outlook For 2025 Remains Uncertain

While AMD’s performance in the data center GPU segment looks like it will be strong this year, and will likely reach a fairly high run-rate exiting the year, I remain concerned about AMD’s ability to sustain its growth trajectory in 2025. Although the MI300 backlog had grown quite rapidly in late 2023 and early 2024, new orders (at least for 2024) seem to have dried up over the last two quarters. In both Q1 and Q2, the order backlog has only grown by about $500 million.

I had downgraded AMD after Q1 based on the slow backlog growth, and I had discussed my worries in more detail in my most recent article about AMD as well. Specifically, I am concerned that AMD’s backlog growth has slowed because it is struggling to compete against Nvidia (NVDA). NVDA has been making rather relentless progress on both the hardware and software fronts, especially as it moves from a two-year to a one-year launch cycle for GPUs. Readers can see my detailed analysis of Nvidia’s strong technological progress here.

AMD CEO Lisa Su was directly asked about the outlook for 2025 on the earnings call. Here was her response:

As we go into next year, I mean one of the important things that we announced at Computex, was increasing and expanding our road map. I think we feel really good about our road map. We are on track to launch MI325 later this year. And then next year, our MI350 series, which will be very competitive with Blackwell solutions. And then we’re well on our way to our CDNA Next as well. So I think, overall we remain quite bullish on the overall AI market. I think the market continues to need more compute. And we also feel very good that our hardware and software solutions are getting good traction, and we are continuing to expand that pipeline.

At least to me, Lisa Su’s response seems a bit noncommittal. She seems confident that the overall market for data center GPUs will grow, which could lead to growth for AMD’s MI300/MI325 sales as well. But at the same time, it does seem like AMD is struggling to keep up with Nvidia’s offerings.

I say this because Lisa Su stated that the MI350 (which will presumably launch toward the end of 2025) will be competitive with Nvidia’s Blackwell lineup (which is launching this year, with Nvidia expecting “a lot of Blackwell revenue this year”). Nvidia is expected to launch Blackwell Ultra GPUs in 2025, and its future-generation Rubin GPUs should also be in production toward the end of 2025. Therefore, to me, it seems increasingly likely that AMD will continue to lag Nvidia in terms of technology through 2025. If so, it is possible that AMD’s strong growth trajectory in data center GPUs could slow in the coming year.

On a more positive note, though, management continues to see increased adoption by some large customers as well as a good engagement pipeline. Specifically, it appears that Microsoft is increasing adoption of AMD GPUs for both internal workloads and cloud customers:

Microsoft expanded their use of MI300X Accelerators to power GPT-4 Turbo and multiple co-pilot services including Microsoft 365 Chat, Word, and Teams. Microsoft also became the first large hyperscaler to announce general availability of public MI300X instances in the quarter.

On the customer pipeline front, management provided the following update:

Our enterprise and Cloud AI customer pipeline grew in the quarter, and we are working very closely with our system and cloud partners to ramp availability of MI300 solutions to address growing customer demand. Dell, HPE, Lenovo, and SuperMicro all have Instinct platforms in production, and multiple hyperscale and tier-2 cloud providers are on track to launch MI300 instances this quarter.

Overall, I would say that things could go either way for AMD in 2025. On one hand, the strong performance in terms of customer engagements and deployments is heartening for investors. On the other hand, orders for the MI300 in 2024 seem to have dried up, which continues to be an ominous sign, especially as Nvidia pursues a very aggressive hardware and software roadmap.

Conclusion

In my most recent article on AMD, I had expressed the following overall view about AMD:

AMD remains one of the more difficult-to-predict AI plays on the market today. Nvidia is being very aggressive with introducing new products on both the hardware and software fronts, which makes it difficult to forecast how well second-place AMD can do in the coming quarters and years. It is possible that AMD could carve out a decent niche for itself in the data center GPU business. But on the other hand, it is possible that Nvidia could make it very difficult for AMD to hang on to more than a foothold.

Based on what we have learned from the Q2 earnings and conference call, I remain uncertain about AMD’s data center GPU prospects in 2025 (and beyond). AMD’s performance so far this year, as well as for the remainder of the year, is quite impressive. Run-rate data center GPU revenue of almost $7 billion exiting the year is a significant accomplishment, especially considering that AMD was almost irrelevant in this segment until the launch of the MI300. But the slowdown in new orders for 2024, along with a seemingly very difficult to overcome technology gap against Nvidia, means that I remain cautious about AMD going forward. As a result, overall, I maintain a neutral Hold rating for AMD stock.

---

Analyst’s Disclosure: I/we have a beneficial long position in the shares of NVDA either through stock ownership, options, or other derivatives. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.

Seeking Alpha’s Disclosure: Past performance is no guarantee of future results. No recommendation or advice is being given as to whether any investment is suitable for a particular investor. Any views or opinions expressed above may not reflect those of Seeking Alpha as a whole. Seeking Alpha is not a licensed securities dealer, broker or US investment adviser or investment bank. Our analysts are third party authors that include both professional investors and individual investors who may not be licensed or certified by any institute or regulatory body.

---