Elon Musk unveils $20 billion ‘TeraFab’ chip project to make chips, memory, and package processors all under one roof — targets a terawatt of annual compute

News
By published

The joint Tesla-SpaceX fab will make both terrestrial inference chips and space-hardened processors.

TeraFab
(Image credit: Tesla / SpaceX)

Tesla and SpaceX CEO, Elon Musk, announced Saturday night that his TeraFab semiconductor project will be built on the Tesla campus in eastern Travis County, Austin, Texas, as a joint venture between the two companies.

In a livestream broadcast via X, Musk stated that the facility exists because the global chip industry cannot expand quickly enough to meet his projected demand across AI, robotics, and space computing. "That rate is much less than we'd like," Musk said from the defunct Seaholm Power Plant in downtown Austin. "We either build the TeraFab, or we don't have the chips, and we need the chips, so we build the TeraFab." The project reportedly carries a $20 billion price tag.

The Austin fab will house equipment for logic, memory, packaging, testing, and lithography mask production in a single building. Musk claimed that capability does not exist at any other facility in the world, and that having everything under one roof enables a rapid iteration loop: make a chip, test it, revise the mask, and repeat without shipping wafers between sites.

Article continues below

The facility is expected to produce two types of chips. One will be optimized for edge inference, primarily for Tesla's vehicles and Optimus humanoid robots. The other will be a higher-power chip hardened for the space environment, which Musk says will run hotter than “terrestrial” designs to minimize radiator mass on satellites.

Musk compared the project to the current global output of global AI compute, which he estimated at roughly 20 gigawatts per year. That figure, he said, represents about 2% of his companies’ eventual needs. On the terrestrial side, he projected 100 to 200 gigawatts per year of chip output; the remainder, up to a terawatt, would go to space-based AI compute aboard solar-powered satellites that SpaceX has already petitioned the FCC to launch.

“That's why I think it's probably a hundred to two hundred gigawatts a year of terrestrial chips, and probably on the order of a terawatt of chips in space," noted Musk. "Just because of power constraints on the ground.”

Musk said Tesla, SpaceX, and xAI — which SpaceX acquired in February — will continue buying chips from existing suppliers, including TSMC, Samsung, and Micron, adding that he would like them “to expand as quickly as they can.” He gave no timeline for when the TeraFab would begin producing chips or reach its target output, and while he has previously referenced 2nm as the target process node, he didn’t repeat that figure in the broadcast.

Google Preferred Source

Follow 3DTested on Google News, or add us as a preferred source, to get our latest news, analysis, & reviews in your feeds.

Luke James
Luke James
Contributor
5 Comments Comment from the forums
  • usertests
    I can buy that getting everything into one building may lead to optimizations and cost reductions. But I didn't hear anything about paying the ASML toll or how they can possibly get to 50x production of everyone else combined. Whereas Substrate ("likened to a fraud") at least has the idea of beating ASML at its own game.
    Reply
  • TerryLaze
    Isn't $20billion a fart in the wind compared to the FABs already producing chips in the USA? Not even counting the $100+ intel invested in the last years or the $100+ that tsmc invested in USA FABs. (Those don't produce yet)
    Reply
  • Zaranthos
    I won't be at all surprised if he ends up with a USA based chip fab that outperforms many other fabs. If he can get Tesla bots doing labor he'll eliminate much of the domestic production argument that labor is too costly in the US. He also continually gets underestimated about his ability to do what the so-called experts say is impossible. According to much of the media and experts Tesla was never supposed to achieve mass production goals, SpaceX would fail, X (Twitter) was doomed to fail as well but still dominates the market space by far. Even if he doesn't outperform he probably still saves money in the long run by making chips in house instead of paying someone else to do it.
    Reply
  • usertests
    Zaranthos said:
    I won't be at all surprised if he ends up with a USA based chip fab that outperforms many other fabs. If he can get Tesla bots doing labor he'll eliminate much of the domestic production argument that labor is too costly in the US. He also continually gets underestimated about his ability to do what the so-called experts say is impossible. According to much of the media and experts Tesla was never supposed to achieve mass production goals, SpaceX would fail, X (Twitter) was doomed to fail as well but still dominates the market space by far. Even if he doesn't outperform he probably still saves money in the long run by making chips in house instead of paying someone else to do it.
    That figure, he said, represents about 2% of his companies’ eventual needs.
    If he "eventually" wants 55-60x of current global fab capacity, that seems unlikely.

    SpaceX has done well and the Falcon 9 capability would outlast the company if it suffered financial problems. Tesla, X, and xAI all face big problems. If the AI bubble bursts, the market won't be kind to Terafab and space datacenter plans.

    At some point, risk beats hype. Car companies are hard. A car company transforming into a robotics company is a crazy gamble.
    Reply
  • thestryker
    Who's going to license all of the IP required to make this functional?

    I absolutely believe such a facility can be made, but no company in their right mind is going to license manufacturing nodes or product technology to make it function. That would mean either buying up smaller companies or developing it themselves both of which are expensive and would take a while. Then there's also being beholden to ASML's timelines since they can only make so many EUV machines per year. Lastly they need sufficient support companies which may be tricky at this scale though I imagine money greases this wheel since Samsung already has fabs in the general area so it should just be expanding existing infrastructure as opposed to entirely new.
    Reply