AMD (AMD) Technology Leadership Forum 2026 summary

Market trends and growth outlook

• Server CPU market experiencing rapid growth, with over 50% growth in Q1 and over 70% in Q2, driven by a shift from AI training to inference workloads.

• Forecasts indicate more than 80% growth in server CPU revenue in Q3 and Q4, and at least 70% growth next year, with a long-term view of 2027 server revenue being 20% larger than the entire 2025 server market.

• AI data center business is expected to grow much more than 100% next year, with significant ramping starting in September and continuing through 2028.

• Total addressable market (TAM) for server CPUs updated to $220 billion by 2030, with agentic workloads projected to represent about two-thirds of that TAM.

• Ambition to reach over 50% revenue share of the expanded TAM, supported by deep customer engagements and a robust product roadmap.

Product innovation and roadmap

• Launching 2 nm Venice CPUs, sampling now and shipping broadly across server markets, including Helios AI racks.

• Venice is the first server product to use advanced packaging, supported by a $10 billion investment in the Taiwanese ecosystem for backend capacity.

• Florence CPU lineup planned for 2028, targeting a range from high-performance single-thread SKUs to high core/thread count CPUs for agentic racks.

• Helios platform, including MI450 and Venice CPUs, enters mass production in September, with a rapid revenue ramp expected in Q4 and Q1.

• Significant acquisition of ZT Systems to enhance system-level talent and de-risk design for rack-scale deployments.

Customer demand and supply chain

• Multi-gigawatt commitments secured from OpenAI, Meta, and Anthropic, each targeting roughly a gigawatt deployment next year.

• Additional demand from core customers like Oracle and Microsoft, with ambitions to ramp supply and meet strong customer demand.

• Close collaboration with TSMC and memory partners to secure supply, with advanced planning and accurate forecasting critical for capacity.

• Focused initial ramp with select ODM partners before broader ecosystem rollout to ensure quality and stability.

• Ongoing efforts to increase near-term supply and support rapid growth, with industry-wide adjustments for future capacity.