- Published on
AI Demand Surges: Intel Rapidly Retrofits Multiple CPU Plants from Desktop to Data Center Processors
- Authors
- Name
- DeepHorus
Facing rapidly rising demand for artificial intelligence infrastructure, Intel is shifting part of its CPU manufacturing capacity from client products to data center products. According to a recent Tom's Hardware report, the chip giant plans to ease tight server processor supply through capacity expansion and adjusted process allocation. Intel CEO Lip-Bu Tan said the industry remains in a severe supply shortage, with superintelligence demand as the main driving force.
The consumer PC market is shrinking, but Intel believes AI industry demand will offset the decline in the client market. To this end, the company has reallocated part of its capacity, expanding Fab 24 near Leixlip, Ireland, to produce more Xeon chips. The expansion project is expected to begin production only by the end of 2027.
Earnings data show that Intel's Data Center and AI business generated $7.3 billion in revenue in the third quarter of 2026, up 17% quarter over quarter and nearly 80% year over year. Lip-Bu Tan expressed satisfaction with progress in server supply growth in the third quarter, but acknowledged that the company still needs to address persistent server demand challenges in the coming quarters. The company expects data center CPU demand to remain strong in 2027 to 2028 as AI infrastructure operators deploy more CPU-intensive agents and inference workloads.
To meet Xeon CPU demand in the coming years, one of Intel's immediate priorities is to increase production of Xeon 6 "Granite Rapids" processors, which are manufactured using the Intel 3 process. The company plans to convert part of the manufacturing resources at its Oregon facility from Intel 18A to Intel 3. The relevant production lines were initially used to manufacture Panther Lake desktop processors, but because of weakened PC demand, Intel will prioritize higher-demand data center products. New Intel 3 capacity in Oregon is expected to come into production in the second half of 2027.
In Arizona, Intel is accelerating Intel 18A production at Fab 52 and speeding up equipment installation at the neighboring Fab 62. Fab 62 will become Intel's second high-volume manufacturing fab for 18A, 18A-P, and derivative process technologies. The company has also committed to producing 18A chips while expanding advanced packaging capabilities in New Mexico and Malaysia.
This capacity adjustment reflects diverging demand trends between PCs and data centers. Increased Xeon supply may ease CPU shortages in AI infrastructure, but it also means reduced elasticity in client CPU supply. Intel needs to maintain a balance between expansion cycles and market demand.
On the timeline, new capacity in Ireland and Oregon will be released by the end of 2027 and in the second half of 2027, respectively, while the Arizona 18A expansion is also advancing. Intel expects AI inference and agent workloads to support server CPU demand. The company is preparing for Xeon supply in the coming years through its manufacturing network and packaging layout.
