Intel data-center and AI revenue rises 78% as Xeon supply takes priority

Intel reported $7.3 billion in third-quarter Data Center and AI revenue, up 78% from a year earlier, on October 29. The chipmaker made expanding Xeon server CPU supply an immediate priority. The reported quarter ended September 26. Total revenue rose 21% to $16.5 billion.

Data Center and AI is a business segment covering several kinds of silicon. Intel’s June-quarter filing describes data-center CPUs, AI accelerators, networking products and custom chips for cloud, enterprise, telecommunications and high-performance computing. Intel Products handles design, development, sales and servicing; Intel Foundry and, to a lesser extent, outside manufacturers make those products. The segment definition therefore includes more than standalone AI accelerators. The separate PC segment also serves commercial, retail and industrial edge computing.

Enterprise AI also needs CPUs to run business applications, move data and coordinate agents’ actions. Intel identified those workloads in its Oracle collaboration announcement, describing capacity for existing systems alongside the accelerators serving larger models.

The investor presentation described strong demand across hyperscalers, enterprises and newer cloud providers. Purpose-built silicon revenue more than doubled year over year. Intel said agentic workloads were increasing customer interest across its product roadmap.

The preceding quarter provides a comparison: Data Center and AI revenue was $6.3 billion in Q2, up 59% year over year. Intel’s July 23 release had already described stronger server demand and additional supply investment. It forecast total Q3 revenue of $15.8 billion to $16.8 billion; the reported result landed within that range.

Production and design behind the demand

Next-generation designs also depend on how chips are manufactured and packaged. Intel’s August Hot Chips disclosure described Diamond Rapids using its 18A-P process and three-dimensional packaging. Adaptable compute blocks, a shared memory fabric and flexible input/output connect the system, reflecting Intel’s stated aim of designing for power, cost and deployment constraints.

Those capacity efforts predate this quarter. In July, Intel said it was increasing equipment, cleanroom and substrate investment to support expected growth across products and foundry. The same release described a €5 billion investment to expand manufacturing capacity for Xeon processors built on Intel 3, establishing the earlier investment baseline.

The October 27 plan coordinates roadmaps and capacity across current and future Xeon platforms.

Next-generation hardware remains at a different stage from reported sales. Intel’s August Hot Chips material positioned Diamond Rapids, its next Xeon, as the CPU for orchestration. A separate data-center GPU addresses inference, while client and edge processors handle device-side AI. That earlier architecture disclosure divides work across different kinds of chips.

Planned OCI X14 offerings would use Diamond Rapids for workloads including embedding generation, search reranking and smaller-model inference. The plan retains existing software; no launch date was specified.

Memory shortages were affecting the PC market, Intel’s presentation said. Client Computing and Physical AI revenue fell 5% year over year to $8.1 billion. The foundry segment recorded an approximately $1.9 billion operating loss.

Intel forecast total fourth-quarter revenue of $15 billion to $16 billion. CFO David Zinsner said the PC slowdown was expected to continue through at least the first half of next year, while the company continued disciplined capacity and working-capital investment to support supply.