The tech billionaire has outlined plans for an ambitious chip manufacturing plant that incorporates his various ventures, including Tesla, SpaceX and xAI.
Dubbed “Terafab,” the proposed project would consolidate semiconductor production for his companies’ AI power needs into a single, large-scale facility, centralizing chip design, fabrication, memory production and packaging in one location.
Such an approach would mark a significant departure from the typically highly specialized, globally distributed semiconductor supply chain.
Initial reports said Musk shared his plans on Saturday night at an event in downtown Austin, Texas, though he later confirmed the news on his X account.
“SpaceXAI + Tesla Terafab Project,” he wrote.
“Goal is a trillion watts of compute/year. Most must necessarily go to space, as U.S. electricity is only 0.5TW,” he continued, referring to the number of terawatts, which is equal to 500 gigawatts, or 500 billion watts of power, which he claims represents the entire U.S. electricity supply.
In a follow-up post, Musk elaborated that Terafab would effectively operate as two fabrication plants, each dedicated to a single chip design. One chip type would reportedly be used in Tesla vehicles and Optimus humanoids, while the second would be deployed in SpaceX’s satellites.
This approach, he said, would simplify manufacturing processes and increase throughput.
“A super high production rate allows us to test very quickly what steps can be deleted, simplified or sped up, even after the design is fixed,” he added. “Current fabs are extremely conservative, operating on rigid historical heuristics, which are mostly, but not all, correct.”
No firm timeline for construction or full-scale operation has been disclosed.
The proposal comes at a time when AI compute demand is heating up, while concerns over supply chain dominance mean governments and companies alike are seeking greater control over production.
Musk pitches the project as necessary to meet these ramping demands.