Analysis
Tesla and SpaceX will jointly build "Terafab," a vertically integrated chip factory in Grimes County, Texas, just outside Houston, with an initial investment of $16.8 billion, according to TechCrunch. The companies plan more than 100 million square feet of manufacturing space, and SpaceX has indicated in filings that total spending across a multi-phase build could reach as much as $119 billion.
Terafab is designed to make, package and test advanced logic and memory chips under one roof, supplying processors for Tesla's Optimus humanoid robots and Cybercab robotaxis alongside high-power chips for SpaceX's space-based data centers. The two companies say combined future compute demand between them is forecast to top 1 terawatt. Intel has agreed to contribute to the project, and the facility is expected to employ at least 3,000 people from Grimes and neighboring Brazos County.
“The two companies say combined future compute demand between them is forecast to top 1 terawatt.”
The move puts Musk's companies in the same category as Amazon, Google, Meta and OpenAI, all of which are now building or backing custom silicon rather than depending entirely on Nvidia and merchant foundries. Unlike those efforts, Terafab is aimed squarely at physical products -- robots and rockets -- not primarily at training large language models, a different demand curve than the data-center buildout driving most other chip announcements this year.
The $16.8 billion initial figure is a fraction of the $119 billion SpaceX has floated as the eventual total, and neither company has published a construction timeline or a date for first chips off the line -- both of which are the real tests of whether this is a serious manufacturing commitment or a headline-grabbing groundbreaking. Pulse has previously covered Tesla's chip ambitions, and Terafab is the largest and least-substantiated one yet: no timeline, no first-chips date, and a $119 billion eventual price tag resting entirely on SpaceX's own filings rather than any third-party commitment -- the real risk worth flagging until ground breaks.
