Intel has secured its first external customer for the 18A-P semiconductor manufacturing process, Chief Executive Lip-Bu Tan said at the TomorrowX Summit in Austin, Texas, according to reporting published on November 19. The customer has not been publicly identified. Intel chief technology officer Pushkar Ranade later confirmed that the agreement relates to the 18A-P node, which entered risk production in June. The announcement is an important commercial milestone for Intel Foundry, but the company has not disclosed order volume, pricing, production start dates or the expected revenue contribution.
The agreement follows comments made by Tan during Intel’s third-quarter 2026 earnings call, when he said the company was engaged with some highly strategic customers about its 18A technology. The latest disclosure moves at least one of those discussions into a signed customer relationship. It does not yet establish the scale of the commitment or whether the customer will use the process for a high-volume commercial product. Semiconductor foundry agreements can begin with design engagement and qualification work long before wafer production reaches meaningful commercial volumes.
A customer milestone for Intel’s manufacturing strategy
Intel 18A is a central node in the company’s manufacturing roadmap and one of the final stages in its previously stated plan to provide five process nodes in four years. It combines RibbonFET gate-all-around transistor architecture with PowerVia backside power providey. RibbonFET changes the transistor structure to improve control of current flow, while PowerVia moves power providey to the back of the wafer to separate power routing from signal routing. These technologies aim to improve density, performance and power characteristics, but commercial success depends on yield, reliability, cost and customers’ ability to complete their designs.
The 18A-P variant can extend the 18A platform with performance improvements while remaining compatible with 18A designs, according to Data Center Dynamics. That compatibility can reduce the amount of redesign required for customers moving between the variants, although the actual work depends on the product and the specific process rules. Intel’s chief executive said during the third-quarter call that yields on Intel 3 and Intel 18A had improved, with 18A showing a double-digit rate of improvement. The company also said wafer output on both nodes exceeded expectations.
Intel has not named the external customer. Earlier reports had discussed possible foundry engagements involving major technology companies, but those reports should not be treated as confirmation of the identity of this customer. The current announcement supports only the conclusion that a customer has signed for 18A-P. Without a name, product, volume commitment or manufacturing schedule, investors and suppliers cannot determine how much of Intel’s planned capacity the agreement will use or how quickly it could translate into revenue.
For a foundry, securing a customer is only one stage in the commercial process. The customer must complete design implementation, tape out the chip, validate the silicon and establish that the manufacturing process can provide the required performance and reliability at an acceptable cost. The foundry must then show that yields and throughput are stable enough for volume production. A signed customer can be an important signal of confidence, but it does not remove the execution risks associated with bringing a new process node to market.
Intel’s foundry ambitions also depend on a broader ecosystem of electronic design automation tools, intellectual-property blocks, packaging services, testing capacity and customer support. External chip designers often build their products around a foundry’s process design kit and qualified IP. They need confidence that the process will remain supported and that the supplier can meet providey commitments over the life of a product. Intel’s ability to attract one customer may help build credibility, but it will need repeat engagements and a growing pipeline to establish a sustainable foundry business.
For customers evaluating a provider, the published capacity figure is only one part of the decision; service-level commitments, regional availability and support arrangements will determine whether the capacity can be used for production workloads.
The company has said that 18A-P remains on track for launch by the end of 2026 across its internal product portfolio, while external customer engagement continues to build. The signed agreement provides a tangible data point for that claim. The next milestones to watch are customer disclosure, design progress, qualification results, wafer starts and any evidence of repeat orders. Until those details emerge, the deal should be treated as an early validation of Intel’s process provideing rather than proof that its foundry turnaround is complete.