AMD projects $120B CPU market as agentic AI shifts workload demands

Additional Context

The upward revision of AMD's market forecast aligns with recent shifts in the competitive landscape for x86 and Arm architectures. Per Mercury Research (May 2026), AMD reached a record 46.2% share of x86 server CPU revenue in Q1 2026, gaining significant ground on Intel, whose share fell to 53.8%. This momentum is supported by the rapid adoption of AMD's EPYC Turin and Genoa lines across hyperscale and enterprise environments. At the same time, the rise of custom Arm silicon from cloud providers poses a concurrent threat to traditional x86 dominance, prompting AMD to focus on total cost of ownership rather than instruction set loyalty. Technological milestones are also accelerating. Per TechPowerUp (June 2025), AMD’s upcoming 'Venice' 6th-generation EPYC processors are expected to deliver a 70% multithreaded performance boost over the current 'Turin' generation. These chips will utilize TSMC’s 2nm process and feature PCIe Gen 6 connectivity to double bandwidth for GPUs and networking. This hardware roadmap is designed to support the 'Helios' rack-scale infrastructure, which targets 2900 FP4 dense PFLOPS to compete directly with Nvidia’s upcoming Vera Rubin series. AMD's $4.9 billion acquisition of ZT Systems, which closed in early 2025, has been pivotal in this pivot toward rack-scale AI delivery. Per AMD (April 2025), the company integrated 1,000 systems engineers from ZT Systems to accelerate the deployment of end-to-end infrastructure for clients like Microsoft and Meta. While AMD plans to divest ZT’s manufacturing business to avoid competing with its own OEM partners, the strategic retention of its design teams allows AMD to provide the highly integrated, liquid-cooled solutions demanded by the power-intensive agentic AI era.