Tesla, Inc. Plans $20 Billion Terafab to Secure AI Chip Supply

New semiconductor facility aims to power autonomous vehicles, humanoid robots, and Tesla’s full AI stack

Elon Musk has unveiled Tesla’s ambitious Terafab Project, a vertically integrated semiconductor manufacturing initiative designed to meet the company’s rapidly rising demand for artificial intelligence chips. Scheduled to formally launch on March 21, the project is expected to become one of Tesla’s largest industrial investments, with an estimated cost of $20 billion and a production target of up to 100,000 wafer starts per month.

The proposed facility is expected to combine logic processing, memory production, and advanced packaging under one roof, enabling Tesla to reduce dependence on external chip suppliers such as Samsung Electronics and Taiwan Semiconductor Manufacturing Company. Reports suggest the fab may be developed near Gigafactory Texas, where large-scale site preparation has already been observed.

Tesla’s next-generation AI5 chip, expected to be one of the first outputs of the plant, is being designed on a 2-nanometre process node and is projected to deliver 40–50 times higher compute capability and nine times more memory than the current AI4 platform. These chips will support Full Self-Driving, Cybercab robotaxis, and the Tesla Optimus humanoid robot program, where future chip demand is expected to far exceed current supplier capacity.

If executed successfully, Terafab would position Tesla among a very small group of companies capable of producing advanced AI silicon in-house, extending its vertical integration strategy beyond vehicles into semiconductor manufacturing.