BofA Increases Price Target for AMD and ARM, Boosts CPU TAM to $170bn

Bank of America Securities analyst Vivek Arya published a research note on June 11, 2026, raising price targets for AMD and ARM while sharply expanding the firm's outlook for the global server CPU market. Arya lifted his AMD objective to $560 from $500, and bumped ARM Holdings to $335 from $245, citing agentic AI as the structural driver behind both moves.

The headline call was a near-40% increase in BofA's long-run server CPU total addressable market estimate, now pegged at more than $170 billion by 2030, up from a prior forecast of $125 billion. Arya argued that agentic AI workloads, which involve orchestration, reasoning, and multi-step decision-making, are better suited to CPUs than to accelerators, implying a 37% compound annual growth rate for the segment between 2025 and 2030. He described the emergence of agentic AI as "a powerful demand accelerant that expands the CPU opportunity and lifts both x86 incumbents and ARM challengers".

For AMD specifically, the revised target reflects higher expectations for both CPU and GPU revenue, with the firm pointing to AMD's next-generation Venice processor platform as a near-term catalyst. AMD shares moved up roughly 4% on the session following the note. Intel, which BofA also covers, rose approximately 8%, underscoring how broadly the agentic AI thesis could benefit traditional server CPU suppliers. ARM Holdings retained a Neutral rating despite the target increase, as Arya highlighted the company's longer-term chiplet architecture potential rather than near-term earnings momentum.

The note adds to a broader Wall Street reassessment of CPU demand in the context of AI infrastructure buildouts. While GPU accelerators have dominated capital spending discussions, BofA's analysis suggests that the software layer of agentic AI, which coordinates large numbers of autonomous tasks, could restore CPUs to a more prominent role in data center architecture. Whether AMD and ARM can fully capture that opportunity will depend on execution against an increasingly competitive landscape that includes Intel and custom silicon from hyperscalers.