Intel could benefit from Tesla and SpaceX’s Terafab plans

Terafab is framed as Tesla and SpaceX’s answer to expected shortages in logic, memory and advanced packaging capacity for autonomous vehicles, humanoid robots and space-based AI compute. The project is planned for a 22,000 acre site in Grimes County, Texas, licensed to SpaceX through 2061, with at least $5Bn of investment required by 2030. Phase one is expected to cost $16.8Bn for construction, while the full vision is described as a 10-phase buildout reaching 100 million sqft and one million wafers per month.

Intel’s involvement remains thinly detailed, but Terafab is expected to use Intel 14A for logic wafers, with Intel positioned to earn upfront NRE payments and per-wafer royalties. The modelling places phase-one high-margin income in a broad range, from $0.5Bn to $6Bn per year around 2030, with additional NRE revenue estimated at $0.3 – $0.5Bn. Longer-term upside could include licensing future nodes such as 10A and 7A, EMIB or Foveros advanced packaging, mask-writing demand through IMS Nano, and a possible DRAM/HBM collaboration.

Risks include losing Tesla and SpaceX as direct Intel Foundry customers, exposing 14A IP outside Intel’s own fabs, higher competition for advanced fab workers, tighter tool supply and the possibility that Terafab eventually brings process development in-house. The opportunity depends heavily on contract protections, wafer mix and whether Tesla and SpaceX scale demand enough to justify the full buildout.