Intel is reallocating manufacturing capacity away from consumer PC processors and toward data center CPUs, as the company said demand from the AI industry for anything that can compute or store data now overwhelms a shrinking client market. The disclosure came on Intel's latest earnings call, where executives described a server business that cannot make chips fast enough.
"The industry continues to navigate one of the most significant supply shortages in history, driven by the unprecedented demand for Super Intelligence," said Intel CEO Lip-Bu Tan during the call, according to Tom's Hardware. "I am pleased by the progress we made growing our server supply in Q3. I am mindful about the ongoing challenges we face in the future quarters to keep pace with server demand." The remarks land amid a broader capital shakeout across the AI sector, which you can track in our ongoing AI industry coverage.
Server Revenue Up Nearly 80% Year Over Year
The financial backdrop explains the urgency. Intel's Data Center and AI (DCAI) business generated $7.3 billion in revenue in the third quarter of 2026, up 17% sequentially and nearly 80% year over year. The company expects demand for data center CPUs to remain strong through 2027 and 2028 as AI infrastructure operators deploy more CPU-intensive agentic and inference workloads — workloads that lean heavily on server processors even when GPUs do the model math.
Granite Rapids, Intel's Xeon 6 family built on its Intel 3 process, has already become the company's fastest-ramping server processor family, and management said demand continues to exceed available supply despite substantial production increases planned over the next several quarters.
From Panther Lake to Xeon Wafers
The most concrete symbol of the pivot: Intel is converting manufacturing resources in Oregon from its Intel 18A node to Intel 3. Those resources were originally intended to build Panther Lake, Intel's client processor family — but weaker PC demand has freed them up for higher-demand data center products instead.
"What we are taking offline to put Intel 3 into Oregon is 18A wafers, which we are generally going to build Panther Lake products," said CFO Dave Zinsner on the call. "We see the opportunity with some of that capacity freed up to go produce more data center wafers." The additional Intel 3 capacity in Oregon is expected to enter production in the second half of 2027.
That is not the only expansion. Intel has begun expanding its Fab 24 complex near Leixlip, Ireland, to produce more Xeons on Intel 3, though that capacity will not come online until the end of 2027 unless the schedule is accelerated. Equipment installation at Fab 62 in Arizona has started ahead of schedule, where the company vows to make 18A chips, and Intel is expanding its advanced packaging capacity in New Mexico and Malaysia.
A Shrinking PC Market, A Booming Server One
The trade-off Intel is making explicit is that the consumer PC market is shrinking and will continue to shrink in the coming quarters, while the AI buildout absorbs everything the industry can produce. The company said it is also improving manufacturing yields and reducing cycle times to extract more usable processors from existing capacity, and that Intel 3 manufacturing costs have declined roughly 15% since the first quarter of 2026.
The move is a notable reversal for a company whose identity was built on the client PC. Just as AI data centers have hoarded memory and advanced packaging capacity, they are now pulling CPU wafer allocations away from desktops and laptops — with Intel betting that agentic and inference workloads will keep Xeon demand elevated for years. For the PC market, the message from Intel's earnings call is blunt: when server demand and client demand collide, client loses the wafers.
It also underscores a frequently overlooked part of the AI hardware story. The industry's attention sits on GPUs and custom accelerators, but every AI cluster still needs armies of server CPUs to feed data to accelerators, orchestrate agents, run databases, and serve inference traffic — the CPU-intensive workloads Intel cited on its call. A single AI data center cannot function on accelerators alone, which is why demand for server processors can exceed supply even in a year when desktop sales are soft. Intel's calculation is that being short on Xeons costs it more revenue today than being short on Panther Lake, and the capacity moves it announced are the physical expression of that bet.
---
Stay Ahead of AIGet the latest AI news, analysis, and breakthroughs — all in one place.
Read more AI news →
