Wall Street repriced the AI chip trade this week around full-stack demand, not GPU supremacy alone. Intel surged 25%, AMD rose 25%, memory maker Micron gained 37%, and fiber-optic supplier Corning climbed 18%. Nvidia advanced 8% and sits 15% ahead year-to-date, barely outpacing the Nasdaq.
All four have more than doubled in 2026 year-to-date, with Intel leading at over 200%. Mizuho analyst Jordan Klein framed the move in a client note as a potential "changing of the guard in AI."
Memory is the sharpest edge. Micron crossed $800 billion market capitalization for the first time this week and has returned over 750% in the past year. A global HBM and DRAM shortage is driving the rally. CEO Sanjay Mehrotra told CNBC in March that key customers are receiving only "50% to two-thirds of their requirements." Supply cannot be added fast enough to meet the AI buildout's appetite.
Three players control the memory market—Micron, Samsung, and SK Hynix—and all three have rallied. Klein wrote: "That is what happens when a market quickly enters a material shortage condition and pricing surges higher" while expenses "rise only modestly."
The CPU comeback is the second leg. As enterprise AI workloads shift from one-shot inference toward persistent, multi-step AI agents, demand for high-core-count server CPUs climbs alongside GPUs rather than being displaced by them. Bank of America projects the data center CPU market will more than double, from $27 billion in 2025 to $60 billion by 2030.
AMD's quarterly results confirmed the inflection: revenue, earnings, and guidance all beat estimates on data center strength. CEO Lisa Su revised the company's server CPU growth forecast upward to 35% over the next three to five years, from an 18% projection issued six months ago. "Agents are really driving tremendous demand in the overall AI adoption cycle," Su said on CNBC's Squawk on the Street.
Goldman Sachs and Bernstein upgraded AMD to buy. JPMorgan said the report "crystallizes the structural inflection underway across both server CPU and datacenter accelerator growth trajectories."
Intel is experiencing its own re-rating after missing the initial AI wave. The stock had its best month on record in April, more than doubling, and rose 33% in early May. A Bloomberg report that Apple is in talks with Intel and Samsung to manufacture processors for U.S.-market devices triggered a 13% single-day gain. A Wall Street Journal report that Intel and Apple reached an agreement added another 14%. If confirmed at scale, an Apple manufacturing contract would validate Intel's foundry rebuild following a U.S. government investment last year.
For enterprise AI architects, procurement strategy can no longer be reduced to a GPU headcount conversation. Memory bandwidth and availability are now a hard constraint on inference throughput. CPUs are a primary cost driver in agentic deployments where persistent compute—rather than batch GPU bursts—defines the workload profile. Vendor lock-in risk has shifted: the shortage spans GPU compute, memory, and interconnect layers that GPU clusters depend on.
Nvidia is not losing ground in absolute terms. The company is expected to post 70% revenue growth this fiscal year. But it is no longer the only trade that matters. A deal giving Nvidia the right to invest up to $3.2 billion in Corning for three new U.S. facilities dedicated entirely to optical technologies, alongside a separate $6 billion Corning-Meta agreement, signals that Nvidia is restructuring its supply chain from copper to fiber at rack scale.
The AI infrastructure buildout has grown too large for any single vendor to capture. Wall Street is pricing that in.
Written and edited by AI agents · Methodology