Intel has announced that it will redirect part of the manufacturing capacity it had planned for PC chips toward expanding production of server CPUs, where demand is growing. On its third-quarter earnings call on October 29, 2026, the company said it would cut some of the Intel 18A capacity originally planned for manufacturing the Panther Lake PC processor, and would expand production of the Granite Rapids Xeon server processor on Intel 3 in Oregon. According to the transcript of the earnings call published by Benzinga, Intel is aiming to begin Intel 3 production in Oregon in the second half of 2027.
The plan does not immediately increase production of next-generation Xeon chips built on Intel 18A. Granite Rapids is an Intel 3 product that is already shipping, and it differs in both manufacturing process and supply stage from the new-generation server CPUs that use Intel 18A or 18A-P. In addition, even if wafer capacity grows, finished-product shipments will not increase if the substrates needed for advanced packaging are in short supply. Intel's plan is a reallocation of production capacity to respond to shifting PC and server demand, and it also highlights several constraints that remain in the semiconductor supply chain.
Reallocating Intel 18A Capacity for Panther Lake to Intel 3
Intel Chief Financial Officer Dave Zinsner explained the capacity reallocation. Asked during the earnings call's Q&A whether the company could take advantage of the PC market slowdown to expand server capacity, he said Intel would reduce some of the Intel 18A wafer capacity it had planned mainly for Panther Lake, in order to begin Intel 3 production in Oregon. Intel intends to use the resulting room to increase production of data center products, where demand is strong.
Panther Lake is a PC processor, while Granite Rapids is a Xeon product for data centers. Intel 18A and Intel 3 are also different semiconductor manufacturing processes. The change is a revision of how factory capacity is allocated across processes and products; it does not mean that Intel 18A wafers that have already been manufactured will be reworked for Granite Rapids. By shifting part of the capacity planned for PCs to servers, Intel aims to increase the supply of existing Xeon products.
On timing, CEO Lip-Bu Tan said the company plans to begin Intel 3 production in Oregon in the second half of 2027. Granite Rapids itself is already in volume production and shipping, and output is also being expanded at the Ireland fab. The second half of 2027 is therefore a target for bringing additional capacity online in Oregon, not the date of Granite Rapids' first shipments.
Intel has not disclosed how many monthly Intel 18A wafers, or what share of capacity, will be cut in this reallocation, nor the specific makeup of the equipment being repurposed. It is also unclear which Panther Lake products will be affected by supply, or to what extent. What can be confirmed at this point is limited to the policy of shifting some capacity to servers and the target for starting Intel 3 production in Oregon.
PC Market Slows While Wildcat Lake Faces Shortages
Behind the decision to reallocate capacity are sharply different demand trends in PCs and data centers. According to Intel's third-quarter 2026 results, revenue in the Data Center and AI (DCAI) segment was $7.3 billion, up 17% from the previous quarter and up 78% from a year earlier. Meanwhile, revenue in the Client Computing and Physical AI Group (CCPG), which handles PC products and others, was $8.1 billion, down 9% from the previous quarter and down 5% from a year earlier.
In absolute terms, CCPG revenue still exceeds DCAI, but there is a wide gap in growth rates. In data centers, CPU supply cannot keep up with demand driven by the expansion of AI-related workloads, while in PCs, customer orders have fallen as the market slows. This difference in demand led to the decision to direct limited manufacturing capacity toward the products with stronger demand. However, the actual number of wafers to be reallocated cannot be calculated from segment revenue alone.
In its official prepared remarks for the earnings call, Intel cited rising memory prices and shortages, along with limited corporate IT budgets, as the main factors behind the PC market slowdown. Higher memory procurement costs have put upward pressure on PC prices, and PC makers are adjusting inventory in line with sales trends, the company said. Intel forecasts that the PC market will shrink by about 12% in unit shipments in 2026 and decline by about 10% again in 2027. These are Intel's market projections, not confirmed statistics.
That said, a slowdown in the overall PC market does not mean demand has weakened for every product. According to Intel, demand for Wildcat Lake, an Intel 18A CPU optimized for lower-priced devices, was strong enough that available supply sold out in the third quarter, even though shipment volume rose about 70% from the previous quarter. Panther Lake's volume ramp is also progressing smoothly, and Zinsner referred on the call to relatively solid demand. A shrinking overall market and supply shortages for certain products can occur at the same time.
In other words, this reallocation of manufacturing capacity does not mean PC products have uniformly stopped selling. Intel expects the PC market to slow but still faces supply that cannot keep up for individual products. In that context, it decided to allocate more capacity to data center CPUs, where demand is especially strong.
The Increase Is for Intel 3 Granite Rapids, Not Next-Generation Xeon
What matters in this plan is to distinguish accurately between the Xeon generations and the manufacturing processes involved. Intel is currently expanding supply of Granite Rapids, which uses Intel 3, while also rolling out Clearwater Forest, which uses Intel 18A, and Diamond Rapids, which uses Intel 18A-P. However, the capacity being added in Oregon this time is for Intel 3, which is separate from any production increase plans for Xeon products on 18A or 18A-P.
The supply status of each product as presented on the October 29 earnings call is summarized below.
| Product | Process | Supply status as of October 29, 2026 | Relationship to this reallocation |
|---|---|---|---|
| Granite Rapids | Intel 3 | Shipping. The largest product family in Intel's server lineup | Target of this increase. Target to start Intel 3 production in Oregon is the second half of 2027 |
| Clearwater Forest | Intel 18A | Paid shipments have begun | A Xeon on Intel 18A. Separate from this Intel 3 expansion plan |
| Diamond Rapids | Intel 18A-P | Sampling. Volume ramp planned for 2027 | A next-generation Xeon on Intel 18A-P. Separate from this Intel 3 expansion plan |
The manufacturing process and supply status of each product are based on Intel's official earnings materials, and the additional production plan in Oregon reflects what was stated in the earnings call Q&A. The table compares the supply stage of each product; it does not indicate differences in manufacturing capacity, performance, or yield. The 2027 production plans are current targets, not confirmed start dates.
Granite Rapids is the current server product manufactured on Intel 3 and is the largest product family in Intel's Xeon lineup. Intel is increasing production at its Ireland fab, and the new Oregon plan would expand that capacity further. Clearwater Forest, meanwhile, is the first server product to use Intel 18A, and paid shipments have begun, with demand expected in areas such as high-density rack-scale systems for AI agents. Diamond Rapids is a next-generation product using Intel 18A-P; it is currently at the sampling stage, with a volume ramp targeted for 2027.
Intel is thus pursuing both the expansion of supply for existing products and the ramp of Xeon chips on next-generation processes. The explanation that Intel 18A capacity is being redirected to Xeon could be taken to mean that Intel 18A Xeon production is being increased, but the plan actually expands Intel 3 manufacturing capacity. Confusing the manufacturing process and supply stage would lead to misunderstanding not only which product is being increased, but also when the supply expansion will begin.
Even With More Wafers, a Shortage of EMIB Substrates Constrains Shipments
Even if Intel expands server CPU production capacity, finished-product shipments will not necessarily increase as a result. The target for starting Intel 3 production in Oregon is the second half of 2027, but the specific scale of the increase and its effect on delivery times for individual products have not been disclosed. Semiconductor manufacturing requires not only wafer fabrication but also back-end processes to assemble the finished chips, so additional wafer capacity does not translate directly into higher shipment volumes.
A challenge Intel particularly points to is a shortage of the organic substrates needed for advanced packaging using EMIB (Embedded Multi-die Interconnect Bridge). EMIB is a packaging technology for connecting multiple semiconductor dies at high density, and assembling the chips requires compatible substrates. According to Intel, supply of such advanced organic substrates is short across the industry, and this is a constraint on expanding finished-product supply even if Intel 3 wafer output is increased.
Zinsner also explained in the earnings call Q&A that supply of server products is constrained by both wafers and substrates. For the Granite Rapids production increase to succeed, therefore, the supply of substrates needed for packaging must grow in parallel with securing manufacturing capacity in Oregon. Even if wafer output rises, a shortage of materials required in the back-end process will limit the number of CPUs that can be delivered to customers.
On the manufacturing processes themselves, Intel says improvements are progressing. In the third quarter, yields improved on both Intel 3 and Intel 18A, and Intel 18A yields reached the level the company considers the industry standard earlier than it had anticipated. Intel reported that manufacturing costs improved by about 15% on Intel 3 compared with the first quarter, and fell by about 40% on Intel 18A over the same period. These are Intel's own assessments, but at the very least they provide no basis for concluding that this capacity reallocation stems from technical problems with Intel 18A or a failure in volume production. Changing capacity allocation in response to demand is a separate matter from evaluating process yields.
Intel also said that an incident at its Israel fab will affect server product shipments in the fourth quarter. However, the supply constrained by this incident is mainly for server products using the 10nm process, which is different from the plan to expand Granite Rapids production on Intel 3. Supply problems are arising across multiple manufacturing processes, and it is necessary to distinguish the cause of each and the products affected.
Intel's approach is to use the spare manufacturing capacity created by the PC market slowdown for data center CPUs, where demand is growing. However, to judge whether supply will actually expand, it will be necessary to check not only the start of Intel 3 production in Oregon planned for the second half of 2027, but also improvements in EMIB substrate supply and how these are reflected in Granite Rapids shipment volumes and delivery times. Only when wafer output and back-end processing capacity both expand will the reallocation of manufacturing capacity help ease the Xeon supply shortage.