Intel warns supply cannot meet surge in AI server CPU demand

Intel said it could not keep pace with a sudden surge in demand for traditional server central processing units used alongside AI accelerators, warning investors that first-quarter revenue and profit would fall short of expectations and rattling its stock.

The company forecast first-quarter revenue between $11.7 billion and $12.7 billion, below the market consensus near $12.5 billion and implying a materially weaker start to the year. Intel said it expects to break even on adjusted earnings per share in the quarter, versus a Visible Alpha consensus that had looked for roughly a five cent profit. Shares plunged in after-hours trading, falling as much as 13 percent in some snapshots while other late-session measures showed declines of roughly 6 to 7 percent and one session saw a fall to $47.60.

Executives framed the shortfall as a supply problem rather than a demand weakness. CEO Lip-Bu Tan said on the company’s conference call, “In the short term, I'm disappointed that we are not able to fully meet the demand in our markets.” CFO David Zinsner added that Intel “expects our available supply to be at its lowest level in Q1 before improving in Q2 and beyond,” and acknowledged the company had been “caught off guard” by surging CPU orders from hyperscalers and other large data-center builders. Zinsner also noted that Data Center and AI revenue showed strong sequential growth and said that “revenue would have been meaningfully higher if we had more supply.” Executives said they were working to “squeeze out as much supply as possible.”

Those supply limits arose even as Intel reported stronger-than-expected fourth-quarter results. The company posted adjusted earnings of $0.15 per share on revenue of $13.67 billion, both beating analysts’ estimates. Intel said its Data Center and AI segment grew roughly 9 percent year over year to about $4.7 billion, marking the fastest sequential growth for that business this decade. At the same time, its Client Computing Group, the PC chip business, saw revenue decline about 7 percent to $8.2 billion, and executives said constrained CPU output forced the company to prioritize high-end and mid-range processors over low-end models.

Intel described factory utilization as high, but cited technical and supply-chain issues that limited the number of shippable parts. Yields on its advanced 18A process node were low, with only a small percentage of chips meeting shipping standards, though the company said yields were improving month to month. Separately, a global memory shortage has pushed up memory prices, increasing PC costs and adding pressure on Intel’s margins.

The imbalance highlights a shifting dynamic in data-center builds. Rapid deployments of GPUs for AI by cloud and hyperscale customers have driven unexpectedly large demand for companion server CPUs, exposing the limits of Intel’s current production footprint and process-node yields. That pressure arrives against ongoing competitive challenges in PCs from AMD and a growing ecosystem of Arm-based designs.

Analysts noted that at the midpoint of Intel’s guidance, first-quarter revenue would represent an approximately 11 percent sequential decline and a 3.9 percent year-over-year drop. Intel told investors it expects supply tightness to be most acute in the first quarter and to improve in the second quarter and beyond, setting the stage for a critical operational stretch as the company works to convert strong AI-driven demand into profitable, sustained growth.