Tesla is not actually going to build its own chip fab. Intel is going to do it for them.
That’s the real takeaway from Intel’s announcement today that it is joining the “Terafab” project alongside Tesla, SpaceX, and xAI — the $25 billion Austin chip factory Elon Musk unveiled with great fanfare last month.
Intel made the announcement on X this morning:
“Intel is proud to join the Terafab project with @SpaceX, @xAI, and @Tesla to help refactor silicon fab technology. Our ability to design, fabricate, and package ultra-high-performance chips at scale will help accelerate Terafab’s aim to produce 1 TW/year of compute.”
Read that sentence again. Intel says its ability to design, fabricate, and package chips is what will make Terafab work. That’s not a supplier relationship. That’s Intel running the fab.
A “Tesla fab” that Tesla isn’t building
When Musk unveiled Terafab last month, he pitched it as a vertically integrated mega-facility that would combine design, lithography, fabrication, memory, advanced packaging, and test under one roof on the north campus of Giga Texas. The headline number was a terawatt of annual compute. The subtext was that Tesla, SpaceX, and xAI were going to muscle into semiconductor manufacturing the same way Tesla muscled into batteries.
We were skeptical at the time, and for good reason. Leading-edge chip fabs take TSMC, Samsung, and Intel a decade and tens of billions of dollars each to stand up. Tesla has never made a wafer in its life. The idea that a car company, a rocket company, and an AI startup were going to stand up a competitive sub-2nm process from scratch was, to put it charitably, ambitious.
Today’s Intel announcement confirms the more realistic version of this story: Terafab is a capacity deal dressed up as a Tesla moonshot. Intel brings the process technology, the equipment expertise, and the packaging — the things that actually make a fab work. Tesla, SpaceX, and xAI bring the demand and, presumably, a big chunk of the capital.
Why Intel needs this more than Tesla does
For Intel, this is exactly the kind of anchor customer the Intel Foundry business has been trying to land since Pat Gelsinger first pitched IDM 2.0. Intel’s 18A process node is finally in volume production, and the foundry unit desperately needs marquee logos to justify the capex. Tesla, which already has a $16.5 billion AI6 deal with Samsung and a TSMC relationship for AI5, adds a third foundry partner and a massive, captive buyer for U.S.-made silicon.
It also slots neatly into the geopolitical story Intel has been telling Washington: advanced American chips for American AI and American EVs, built on American soil, with CHIPS Act support.
For Tesla, the framing is harder to square with what Musk has been saying. Just last month, Musk was telling investors that Tesla would “have to” build its own fab because even the best-case scenario from TSMC and Samsung wasn’t enough to supply the AI5, AI6, and Optimus chips the company needs.
Now, it’s starting to look like Tesla and SpaceX will finance and own a fab built by Intel.
Electrek’s Take
We said when Terafab was announced that it reeked of desperation, and this announcement doesn’t change that read, it reinforces it. A real Tesla fab would be a decade-long, $50-billion-plus commitment with a completely new engineering organization. What’s actually happening is that Tesla is co-anchoring an Intel Foundry expansion in Austin and getting to call it “Terafab” in the press releases.
To be clear, that’s probably the right move. Intel 18A is real, Intel’s packaging is genuinely world-class, and the last thing Tesla should be doing right now is spreading its resources even more amid a talent exodus. If Musk has quietly accepted that Tesla’s “gigantic chip fab” is really going to be an Intel fab with Tesla’s name on the purchase order, that’s a win for shareholders and a win for Intel Foundry.
It’s just not the story Musk told a month ago. Tesla isn’t going to build its own chip fab. Intel is going to do it for them,
FTC: We use income earning auto affiliate links. More.
