AMD Surges 19% to Record $421 as CPU Market Forecast Doubles to $120B

Advanced Micro Devices (AMD) ranked second in the U.S. market turnover for the day after rising 18.64% to close at $421.39 on Wednesday, a record high, on trading volume of $36 billion. AMD outperformed the Philadelphia Semiconductor Index and the Nasdaq 100 on the session, increasing its market capitalization to almost $687 billion and extending its year-to-date gain to almost 97%.

The catalyst was not merely a quarterly beat. It was a forecast revision that reframes the entire AI infrastructure narrative around CPUs — and by extension, around AMD.

The Number That Moved the Market: Server CPU TAM Doubled in Six Months

AMD's Q1 2026 results were strong — data centre revenue rose 57% year-over-year to $5.8 billion, providing concrete evidence of expanding demand. But the earnings beat was not what drove an 18.64% single-session move.

The trigger was AMD's revision of its server CPU addressable market forecast, raising the projected annual growth rate from 18–20% (estimated in November 2025) to more than 35%, and projecting the total market will exceed $120 billion by 2030 — more than double the $60 billion figure projected just six months earlier.

CEO Lisa Su explained the revision to CNBC with unusual directness: conversations with large customers over the past 90 days revealed that AI workloads are changing faster than AMD had modelled. "The practical application value of AI is increasing at a faster pace than it did six months ago," she said. The revision reflects observed demand, not extrapolation.

Why AI Agents Are the CPU Story — Not Just the GPU Story

AI agents—autonomous systems that manage multi-step processes for enterprise users across databases, apps, workflows, and communication systems—are the analytical insight behind AMD's reevaluation. "In the future, each person may use two, three, or even five to ten AI agents simultaneously," Su explained the technique. Data centers will require additional general-purpose CPU processing resources in addition to AI accelerators since these agents will handle duties for consumers and workers.

The implication is structural. GPU-based inference handles the AI computation within each agent; CPU resources handle coordination, scheduling, data access, and enterprise system integration across agents. As agent deployments scale — from experimental to production, from single users to entire organisations — the ratio of CPU demand to GPU demand approaches 1:1 rather than the heavily GPU-weighted ratio that characterised the first phase of the AI buildout.

This dynamic reframes AMD's competitive positioning. The AI infrastructure story of 2023–2025 was concentrated almost entirely on GPU demand and Nvidia's dominance in that segment. AMD's Q1 statement introduces a parallel track: a $120 billion server CPU market growing at 35%+ annually, in which AMD's EPYC processors are a primary beneficiary — not as a consolation prize to Nvidia's GPU business, but as a distinct and structurally necessary AI infrastructure layer.

Wall Street's Response: 20+ Target Upgrades, $579 High-End Call

At least 20 brokerages raised their AMD price targets following the earnings call. Evercore ISI set the most aggressive new target at $579 — a level that implies approximately 37% additional upside from Wednesday's record close. The upgrade framing from multiple desks shifted explicitly from characterising AMD as a "Nvidia follower" to positioning it as a beneficiary of both the CPU and accelerator segments of AI infrastructure expansion.

That repositioning is the market's real acknowledgement of what Su's forecast revision represents. If the server CPU market reaches $120 billion by 2030 at 35%+ annual growth, and AMD maintains or expands its market share in server CPUs against Intel, the revenue contribution from that segment alone would justify a significant valuation re-rating independent of AMD's GPU competitive dynamics with Nvidia.

The Intel Parallel — and the Competitive Stakes

The prediction made by AMD also subtly speaks to Intel's potential. Citi and KeyBanc came to similar conclusions regarding CPU demand in the AI age after Intel's own Q1 2026 beat, which was fueled by 22% year-over-year growth in its data center and AI business. Two CEOs from rival firms have independently noted the same structural shift: Lip-Bu Tan's remark that "the CPU is reinserting itself as the indispensable foundation of the AI era" and Lisa Su's claim that GPU-to-CPU demand ratios are getting close to 1:1.

The server CPU market at $120 billion by 2030 is large enough for both AMD and Intel to grow revenue substantially. The competition between them — in process node technology, product roadmap execution, and enterprise customer relationships — will determine market share within that expanding pool.

Investor Takeaway

AMD's 18.64% single-session move on a forecast revision rather than a pure earnings beat is the market recognising that the CPU is no longer a mature, slow-growth category — it is an AI infrastructure essential with a growth rate that rivals or exceeds GPU demand in the agentic AI era. $421 is the record close; $579 is Evercore's bull case.

The path between them requires AMD to continue executing on data centre CPU growth, maintain its EPYC server market share against Intel, and demonstrate in subsequent quarters that the $120 billion TAM forecast translates into durable, accelerating revenue. Wednesday's data — 57% year-over-year data centre growth — provides the first quarter of evidence. Su said AI is still in a "very early" stage. If that's true, AMD's repricing may only be beginning.