AMD's Lisa Su on AI & Computing Super Cycle

The recent Goldman Sachs Communacopia + Technology Conference gave AMD's Chair and CEO, Dr. Lisa Su, a chance to share some really important stuff about the company's plans, where it stands in the market, and what's coming up. This analysis breaks down the most important points from the conference and gives a good look at where AMD is at right now and where it's headed.

AI: The Driving Force Behind AMD's Growth

Artificial Intelligence (AI) has become the cornerstone of AMD's strategy, with Dr. Su emphasizing its importance right from the start of the conference. The company's focus on AI has yielded impressive results, particularly in the data center GPU business.

"It's been a remarkable year. I think we're all in technology, we're moving faster than ever. And in the last year, I mean, if you look at what we've been able to do, we've launched MI300X in December. It's had just tremendous customer traction and customer -- customers have been really excited about it."

The MI300X, launched in December, has garnered significant attention from major players in the tech industry. Dr. Su highlighted the adoption of MI300 by large hyperscalers, including Microsoft, Meta, and Oracle, as well as OEM and ODM partners. This widespread adoption underscores the product's competitiveness and AMD's growing influence in the AI hardware market.

Software Progress: A Key Differentiator

One of the most significant achievements for AMD in the AI space has been its progress on the software front. The ROCm software stack has seen substantial improvements, addressing previous concerns about the ease of entering the AMD ecosystem.

"When I think about, though, what do I believe we've done the best over the last, let's call it, 9 months or so, it's really been the progress on software. That was always a big question around how hard is it to get people into the AMD ecosystem. And we've just made tremendous progress with our overall ROCm software stack."

This progress has enabled AMD to work with some of the most challenging and largest models, sometimes achieving performance that surpasses the competition. This development is crucial as it positions AMD as a strong alternative in the AI hardware market, potentially challenging the dominance of established players.

Strategic Acquisitions to Bolster AI Capabilities

AMD's commitment to AI is further evidenced by its recent strategic acquisitions. The company has made moves to strengthen its software capabilities and infrastructure offerings:

1. Acquisition of Silo AI: This leading AI software company acquisition bolsters AMD's AI software capabilities.
2. Acquisition of ZT Systems: This move enhances AMD's ability to provide rack-scale infrastructure, which is crucial for large-scale AI deployments.

These acquisitions demonstrate AMD's holistic approach to AI, covering not just hardware but also software and infrastructure aspects. This end-to-end strategy could provide AMD with a competitive edge in the rapidly evolving AI market.

AMD Acquires ZT Systems: Data Center AI Expansion

AMD’s $4.9 billion acquisition of ZT Systems marks a significant expansion in data center AI capabilities, aiming to compete with Nvidia and capture a larger share of the projected $400 billion market by 2027.

Giro's NewsletterGiro Lino

The Expanding AI Market: AMD's Long-term View

AMD's long-term outlook for the AI market remains bullish. The company had previously forecasted a $400 billion AI accelerator TAM (Total Addressable Market) by 2027, a prediction that initially seemed ambitious but is now viewed as increasingly realistic.

"When we originally talked about a $400 billion TAM in the 2027 time frame, I believe many thought that, that was high. And actually, I think as time has passed over this last year, I think we're -- we feel very good about that overall TAM."

Dr. Su attributes this confidence to several factors:

1. Early Stage of AI Computing Cycle: The industry is still in the nascent stages of AI adoption, with significant growth potential across various applications.
2. Increasing Compute Demands: Whether for training large language models, inferencing, or fine-tuning, AI workloads consistently demand more compute power.
3. Diverse Market Composition: While GPUs are expected to form the largest portion of this TAM, there's also room for custom silicon solutions.

The company's strategy involves an end-to-end play across different compute elements, positioning AMD to capitalize on various segments of this growing market.

Inference: A Growing Opportunity

While training large AI models has been a primary focus, Dr. Su highlighted the increasing importance of inference in the AI landscape:

"We're also seeing many people have said inference will continue to increase over time. We're certainly seeing that. Training is very, very important, but inference is increasing over time."

This trend towards increased inference workloads presents another growth avenue for AMD. As AI models are deployed in real-world applications, the demand for efficient inference solutions is likely to surge, providing AMD with opportunities to expand its market share.

Server CPU Market: A Turnaround in Sight

While AI has been the headline-grabber, AMD's server CPU business remains a crucial component of its overall strategy. Dr. Su expressed optimism about the server CPU market trends, noting a turnaround after an extended correction period.

"I am pretty happy with some of the server CPU market trends. I think what we've seen is traditional compute is important. So as important as accelerated compute is, there are lots of workloads that run on traditional CPUs."

Key points regarding the server CPU market include:

1. Upgrade Cycle: Customers are beginning to upgrade their infrastructure, driven by the significant benefits in power savings, space efficiency, and overall TCO (Total Cost of Ownership).
2. Strong Adoption of Zen 4: The Genoa family, based on AMD's Zen 4 architecture, has seen robust adoption due to its advanced capabilities.
3. Excitement Around Zen 5: The upcoming Turin cycle, based on Zen 5 architecture, is generating considerable interest among customers.

Cloud vs. Enterprise Adoption

While AMD has seen significant success in the cloud and hyperscale space, enterprise adoption has been comparatively slower. However, Dr. Su noted increasing traction in the enterprise segment:

"As you look at enterprise and some of the, let's call it, the third-party adoption, they've had many other things on their mind, and so they weren't necessarily focused on versus CPU. But at this point, it's all about TCO, and it's all about efficiency."

To accelerate enterprise adoption, AMD has:

1. Increased field application engineer deployment
2. Conducted more complex POCs (Proof of Concepts) for customers
3. Enhanced software support to increase customer comfort

These efforts are paying off, with enterprise adoption showing signs of improvement. Dr. Su emphasized that there's no fundamental technological reason for the disparity between cloud and enterprise market share, suggesting significant growth potential in the enterprise segment.

Custom vs. Merchant Silicon: A Balanced Approach

As more cloud providers develop custom CPU and accelerator programs, questions arise about the future mix of merchant versus custom silicon. Dr. Su provided insights into AMD's strategy in this evolving landscape:

"I find this to be an interesting question because people are always wondering, well, is it going to be X or Y? And I say, look, it's going to be both."

AMD's approach involves:

1. Continued Investment in Competitive Roadmaps: Leveraging economies of scale in architecture, software, yields, and reliability.
2. Partnering with Hyperscalers: Allowing customers to benefit from AMD's scale while offering customization options.
3. Focusing on Performance-Sensitive Applications: Custom silicon is expected to play a role in less performance-sensitive applications or narrow accelerator use cases.

Dr. Su views this not as competition but as an opportunity for closer partnerships, similar to AMD's successful semi-custom business in the gaming console market with Sony and Microsoft.

AI PCs: The Next Growth Driver

The concept of AI PCs has generated significant buzz in the tech industry, and AMD sees this as the beginning of a multi-year cycle. Dr. Su provided insights into AMD's perspective on AI PCs:

"I believe that we are at the start of a multiyear AI PC cycle. So again, you guys are always trying to go a little bit too fast. So we never said AI PCs was a big 2024 phenomena. AI PCs is a start in 2024. But more importantly, it's the most significant innovation that's come to the PC market in definitely the last 10-plus years."

Key points about AI PCs include:

1. Gradual Adoption: While AI PCs are starting in 2024, the significant impact is expected to unfold over several years.
2. Productivity Enhancement: AI capabilities are seen as a natural evolution for PCs as productivity tools.
3. Hardware-Software Co-optimization: Extensive work has been done with partners like Microsoft on initiatives such as Copilot+.
4. Commercial Refresh Cycle Driver: AI PCs are expected to drive a commercial PC refresh cycle in the coming year.

This focus on AI PCs could potentially help AMD improve its position in the commercial PC market, where it has traditionally been underrepresented.

Embedded Business: Signs of Recovery and Synergies

While often overshadowed by discussions about AI and data center products, AMD's Embedded business remains a significant part of its portfolio. Dr. Su provided updates on this segment:

"The Embedded business is a business we don't quite talk about as often as it relates to AMD, but it's a very, very good business for us. When we look at sort of the diversity of customers and the diversity of applications, we continue to believe it's a strong pillar of our overall strategy."

Key points about the Embedded business include:

1. Recovery from Bottom: Q1 marked the bottom for the Embedded business, with improving order patterns observed in Q2 and expected in the second half of the year.
2. Market Variations: Some markets, like aerospace and defense, and test and measurement, are showing stronger recovery than others.
3. Portfolio Synergies: AMD is seeing significant synergies between its embedded FPGA and CPU businesses, with design wins up 40% year-on-year to over $7 billion in the first half of the year.

These synergies are particularly important as they allow AMD to deepen relationships with customers and position itself as a comprehensive solution provider.

Supply Chain and Manufacturing Strategy

The discussion touched upon supply chain challenges and AMD's manufacturing strategy, particularly in light of geopolitical concerns:

"It's clear that we all have to think about sort of resiliency in our supply chain. So COVID certainly taught us that. We continue to look at diversification of the supply chain."

Key points on supply chain and manufacturing include:

1. Ongoing Constraints: Advanced packaging and high-bandwidth memory remain tight, with demand continuing to outpace supply increases.
2. Capacity Increases: AMD has ramped up its ability to service AI revenue in 2024 and plans another significant increase in 2025.
3. Geographical Diversification: Support for initiatives like the CHIPS Act and TSMC's Arizona fab to create more geographic diversity in the supply chain.

While TSMC remains a crucial partner for AMD, the company is actively looking at ways to derisk its supply chain, recognizing this as an industry-wide challenge.

Investment Strategy and Capital Allocation

As AMD navigates numerous growth opportunities, questions arise about balancing investments with profitability and cash flow generation. Dr. Su provided insights into AMD's approach:

"Capital allocation is incredibly important for us, and we do have many more opportunities than – every year, we seem to get more. I think the key principle is we are investing in the business. I mean this is an opportunity for us. I think this AI sort of technology arc is really a once-in-50 years type thing. So we have to invest."

Key points on investment strategy include:

1. Prioritizing Business Investment: Recognizing the unique opportunity presented by the AI revolution.
2. Disciplined Approach: Despite the need for investment, AMD aims to grow OpEx slower than revenue growth.
3. Long-term View: Investments are made with a strategic, long-term perspective.

This approach reflects AMD's commitment to capitalizing on the AI opportunity while maintaining financial discipline.

Conclusion: AMD's Position in the Computing Super Cycle

Dr. Su concluded the conference by emphasizing the broader context of AMD's strategy:

"I think the main thing is, look, this is a computing super cycle, so we should all recognize that. And there is no one player or one architecture that's going to take over. I think this is a case where having the right compute for the right workload and the right application is super important."

This statement encapsulates AMD's strategy:

1. Diversified Portfolio: Offering the best CPU, GPU, FPGA, and semi-custom capabilities.
2. Workload-Specific Solutions: Recognizing that different applications require different computing approaches.
3. Ecosystem Partnership: Positioning AMD as the best computing partner to the broader tech ecosystem.

As the computing landscape continues to evolve, particularly with the rise of AI, AMD's diversified approach and strategic investments position it to capitalize on multiple fronts. The company's progress in AI, coupled with its strong foundation in traditional computing markets, presents a compelling narrative of growth and innovation.

The coming years will be crucial as AMD seeks to translate its technological advancements and strategic positioning into sustained market share gains and financial performance. With a clear vision and a comprehensive strategy spanning hardware, software, and infrastructure, AMD appears well-positioned to play a significant role in shaping the future of computing.