Why Elon Musk is building ‘Tera-fab’: A mega chip factory to fuel his AI ambitions

Elon Musk has announced plans to build a massive semiconductor manufacturing facility, dubbed the “Tera-fab,” in the United States to produce in-house chips for powering AI, robotics, autonomous driving, and space-based computing across his companies, including Tesla, SpaceX, and xAI. The facility will be located near Tesla’s Austin headquarters and is expected to be jointly operated by Tesla and SpaceX. Designed as an advanced fabrication unit, it will be capable of manufacturing and testing a wide range of chips, with a long-term goal of generating up to a terawatt of computing power annually.

Musk’s move comes amid growing concerns over chip shortages and the slow pace of semiconductor supply, especially as demand surges for AI and autonomous technologies. While Tesla currently relies on partners like TSMC, Micron, and Samsung, Musk believes building an in-house facility is essential to meet future needs. The Tera-fab is expected to produce cutting-edge 2-nanometre chips, including edge and inference chips for applications like robotaxis and humanoid robots, as well as high-performance chips for orbital data centres.

A key part of Musk’s vision involves creating a network of orbital data centres, starting with smaller AI-powered satellites capable of delivering 100 kilowatts of computing power, eventually scaling to megawatt levels. These space-based systems are intended to support xAI’s growing computational demands. Musk has even proposed launching up to one million data centre satellites, though the plan has faced opposition and regulatory hurdles. While the ambitious project aligns with industry trends, questions remain about execution, given the massive investment and technological complexity required.