Intel Promises Apple and Terafab a 14A PDK Milestone This Fall

The PDK milestone coming this fall for Intel's 14A chip manufacturing process is, per The New York Times reporting, less a product announcement and more a gate event: the specific technical document that allows prospective customers to begin serious test-chip design work. Intel has already distributed an early PDK 0.5 to potential customers; the next version, expected this fall, is what engineering teams at Apple, AMD, Nvidia, and the Terafab consortium need before they can run real design evaluations against the node.

The 14A process uses High-NA EUV lithography and second-generation RibbonFET transistors, and Intel claims it will deliver a 15-20% performance-per-watt improvement over the preceding 18A process. CEO Lip-Bu Tan has said plainly: "We are going big time into 14A." According to NineScrolls, Intel has also ramped chipmaking equipment orders by more than 50% year-over-year since early 2026, backing the foundry ambitions with real capital.

The prospective customer list is notable. Apple is reportedly considering 14A for lower-end A-series iPhones and M-series chips. AMD and Nvidia are evaluating it for server CPUs and GPUs respectively. And Terafab, a reported $25 billion joint venture between Tesla, SpaceX, and xAI, is planning to use Intel's 14A process for AI chip production. Firm supplier decisions are expected between the second half of 2026 and the first half of 2027, which makes the fall PDK release the pivot point: customers need it before they can begin the engineering evaluations that lead to manufacturing commitments.

The honest caveat is that customer interest and customer commitment are different things. Intel's 18A node entered high-volume manufacturing in October 2025, which gives 14A evaluators some evidence of Intel's yield improvement trajectory, but 14A with High-NA EUV is relatively uncharted territory at production scale. What the reporting does not settle is whether Apple and Terafab are evaluating 14A as a hedge against TSMC concentration risk or as a genuine primary-supplier play, and the two scenarios carry very different revenue implications for Intel. If the fall PDK arrives on schedule and test chips perform, Intel could enter 2027 with anchor customers most observers did not expect it to have.