Elon Musk has unveiled Terafab, a semiconductor manufacturing giant set to become the largest on-and off-Earth. Announced during a livestream, this $20-$25 billion venture is a collaboration between Tesla, SpaceX, and xAI. The first phase involves building a massive chip plant adjacent to Tesla’s Giga Texas in Austin, but the ambitions reach far beyond our planet.
Why is this significant? If Musk delivers, Terafab could disrupt the global chip supply chain. The factory aims to produce 1 million chip wafers monthly using cutting-edge 2-nanometer technology. For context, TSMC-the current leader-plans to reach 140,000 wafers per month by the end of 2026. Terafab’s output would be nearly ten times that, powering Tesla’s vehicles, robots, and a next-generation constellation of AI satellites.
Chips on Earth, Compute in Space
The real twist? Most of Terafab’s computing power won’t stay on Earth. Musk envisions 80% of its chip output running in space aboard SpaceX’s solar-powered AI satellites. Solar energy is about five times stronger in orbit, and cooling is simpler in the vacuum of space. The goal: 100-200 gigawatts of computing power on Earth, and an astonishing 1 terawatt in orbit. To put that in perspective, the entire U.S. electricity demand is roughly 0.5 terawatts.
Currently, compute power in low Earth orbit is minimal, mostly handling basic satellite tasks. China’s “Three-Body Computing Constellation” is the closest comparison, but even its planned 2,800 satellites pale in scale compared to Musk’s vision. If realized, Terafab could open a new frontier for AI and data processing high above the clouds.
Why Build Terafab?
Musk argues that existing chip suppliers-Samsung, TSMC, Micron, and others-can’t scale quickly enough to meet Tesla’s needs. “We either build the Terafab, or we don’t have the chips,” he said. This plant will let Tesla design, manufacture, test, and iterate chips all under one roof-something Musk claims “does not exist anywhere else in the world.”
Speculation suggests that if Terafab succeeds, Tesla could become self-reliant for advanced AI and memory chips, cutting dependence on Asian suppliers and potentially shifting the semiconductor industry’s power dynamics. But it’s a bold leap for three companies with no prior experience building chips at this scale.
Timelines, Tech, and Moonshots
Musk’s timeline remains vague. Tesla’s next-gen AI chip (the AI5) is expected to enter volume production in 2027, but it won’t initially be made at Terafab. The facility itself is likely years from completion. Musk also hinted at a lunar “mass driver” to launch even more compute power into space, though details are scarce and sound more like sci-fi than a concrete plan.
For those tracking Musk’s megaprojects, past promises-like those from Tesla Battery Day 2020-have often slipped deadlines. Still, even partial success with Terafab could reshape chip manufacturing and space-based computing for years ahead.
The bottom line
- Terafab aims to surpass TSMC and deploy most of its compute power in orbit.
- If successful, Tesla could control its chip supply and pioneer AI processing in space.
- Timelines are uncertain, and Musk’s history suggests delays are likely.