The Terafab chip project received a public statement from Intel on October 7, 2026: according to Reuters that day, Intel CEO Lip-Bu Tan told Bloomberg, ahead of an event in Tokyo marking 50 years of Intel's operations in Japan, that Intel will continue to work on Elon Musk's Terafab project. The remark drew attention because just last week Musk confirmed that TSMC had held exploratory discussions with Terafab, and until now Intel had been the project's only publicly named chip partner.
Why TSMC makes Intel nervous
Terafab is a chip manufacturing project in Texas advanced by Tesla and SpaceX, with chips planned for Optimus robots, Cybercabs and space-based data centers. Intel announced it was joining in April 2026, to design, fabricate and package high-performance chips alongside SpaceX, xAI and Tesla, and Musk said on a Tesla earnings call at the time that the plan was to use Intel's 14A process. The project's first stage, in Grimes County, carries a planned investment of more than $16.8 billion.
Last weekend, responding on X to reports that TSMC might take part in Terafab, Musk said only that the talks were still at the discussion stage and something might come of them. Even that understated confirmation knocked Intel shares down more than 4% in premarket trading on October 5, before the stock closed the day down just 0.5%. The market's sensitivity has a concrete reason: in the TSMC involvement structures analysts have outlined, whether TSMC owns and operates a dedicated fab or SpaceX holds a majority stake while TSMC provides technology and operating know-how, neither leaves a role for Intel, and 14A is Intel's most important card in this project.
The weight and the limits of Tan's remark
Tan's statement confirms an intention to participate, not a new contract. Intel has still disclosed no financial terms for Terafab, and neither Intel nor Tesla responded to Reuters requests for comment. Another easily missed detail: SpaceX's May IPO filing described Terafab as a general framework, with specific projects still subject to separate agreements, and neither Tesla nor Intel obligated to remain part of it. Musk has also said previously that chips from TSMC would likely be in addition to those supplied by Intel, not a replacement. For a project aiming at terawatt-scale annual computing capacity, running multiple partners in parallel makes sense on capacity logic; Intel staying in does not automatically mean TSMC stays out.
What to watch next
The real signals are not statements but three follow-ups: whether Terafab's first stage produces a formal process or foundry agreement, whether the TSMC discussions turn into a concrete fab plan, and whether Intel's 14A production ramp can keep up with the project's timetable. For Intel's foundry business, Terafab's value right now is credibility more than revenue; it is the most closely watched external customer endorsement since Tan took over. Losing it would damage the believability of the whole foundry story. By saying publicly in Tokyo, days after the TSMC reports surfaced, that Intel will keep participating, Tan was steadying exactly that expectation.
