Advanced Micro Devices, Inc. (AMD) TD Cowen's 53rd Annual Technology, Media and Telecom Conference (Transcript)

At TD Cowen's 53rd Annual TMT Conference, AMD executives Mark Papermaster and Matt Ramsay discussed the company's focus on scaling AI capabilities and addressing the broader compute landscape. AMD aims to compete in inference first, then training, with upcoming products like the MI350 and MI450, emphasizing total cost of ownership and innovation. The company also highlighted strong revenue share gains in the client (PC) market, driven by ASP increases and leadership in AI-enabled PCs, while managing supply chain dynamics; AMD expects margin expansion in the second half of the year due to a shift in business mix towards enterprise and higher-margin segments.

Analysis

Advanced Micro Devices (AMD) is strategically capitalizing on the burgeoning artificial intelligence sector, with CTO Mark Papermaster highlighting AI's rapid transition from novelty to a critical business driver, creating unprecedented demand for compute power. The company's data center AI accelerator roadmap is central to this strategy, evidenced by the MI300's rapid ascent, achieving approximately $5 billion in revenue in 2024 after launching in late 2023. AMD is pursuing an aggressive product cadence, with the upcoming MI350 family promising a 35x improvement in inferencing performance while maintaining compatibility with existing universal baseboard infrastructure, and the 2026 MI450 family targeting massive-scale training workloads. This progression underscores AMD's focus on first establishing a strong foothold in AI inference, leveraging memory and chiplet technology advantages, before expanding aggressively into the training market. Concurrently, AMD is bolstering its software ecosystem with bi-weekly ROCm updates since Q1 2025 to support hyperscalers and the broader developer community, and enhancing its rack-scale offerings through the ZT Systems engineering team acquisition. In the client segment, AMD reported a robust 68% year-over-year revenue increase in Q1, largely driven by higher average selling prices (ASPs) and leadership in AI-enabled PCs, with management expressing confidence in the sustainability of these ASP gains. The company also anticipates a tailwind from enterprise CPU refresh cycles for its EPYC servers as businesses rebalance IT investments alongside AI buildups. Financially, AMD expects gross margin expansion in the second half of the year, attributed to a favorable product mix, including growth in higher-margin enterprise client and server CPUs, and the removal of approximately $1.5 billion in lower-margin MI308 shipments to China due to export restrictions, even as new large AI deals might be initially dilutive to the overall gross margin percentage but accretive to gross profit dollars.

Key Decisions for Investors

• Investors should closely monitor the launch details and market adoption of the MI350 and subsequent MI400/450 series, as successful execution is critical for AMD to capture significant share in the AI accelerator market and improve segment profitability.
• Evaluate the durability of AMD's strong ASPs and market share momentum in the client PC segment, particularly in relation to AI PC adoption and the anticipated commercial refresh cycle.
• Track AMD's gross margin trajectory in the second half of the year, paying attention to the balance between the dilutive impact of new data center GPU sales and margin improvements from enterprise CPU segments and overall product mix.
• Assess AMD's progress in penetrating the enterprise AI market and the continued maturation and community adoption of its ROCm software platform, which are key to diversifying its AI customer base beyond hyperscalers and competing effectively.