Elon Musk’s Terafab project is a planned semiconductor manufacturing complex designed to produce large amounts of computing hardware for Tesla and SpaceX.

The project targets up to 1 terawatt of computing power per year, with applications including humanoid robots, autonomous vehicles and data centers operating in space.

What Terafab Is Designed To Produce

Terafab was announced as an initiative involving Tesla, SpaceX and xAI, with the goal of addressing the companies’ expected demand for computing hardware. Musk said existing semiconductor production would not provide enough capacity for the computing requirements he expects from future robotics, artificial intelligence and orbital data centers.

The project is planned around two broad categories of chips. AI5 and AI6 chips are intended for applications including Tesla vehicles and Optimus, while the project also identifies a D3 chip for space-based computing.

The facility is intended to combine chip design, fabrication-related work, memory, advanced packaging and testing. Keeping these activities closely connected is intended to provide faster feedback for chip design changes, including modifications to photomasks used during manufacturing.

The Scale Of The Planned Facility

The project is much larger than the initial Austin advanced technology facility. The planned Grimes County complex could eventually reach roughly 100 million square feet, with semiconductor production at that scale requiring substantial land and power infrastructure.

By August 2026, SpaceX and Tesla had identified Grimes County, Texas, as the planned location for the larger Terafab complex. The companies announced an initial investment of $16.8 billion, while later expansion could bring the total investment to $119 billion.

The planned facility is expected to employ at least 3,000 people from Grimes and nearby Brazos County. SpaceX has also said the project will include its own power infrastructure, with natural-gas power plants and large battery arrays planned to supply electricity.

Why Robots And Space Data Centers Need The Chips

Tesla’s Optimus program is expected to require large quantities of computing hardware, while SpaceX’s plans for orbital data centers would require chips capable of operating in space.

Musk has projected demand for chips supporting roughly 100 to 200 gigawatts of computing capacity per year for terrestrial applications, including vehicles and humanoid robots. He has also projected that global humanoid-robot production could eventually reach 1 to 10 billion units annually, with Tesla aiming to produce a significant share.

For orbital computing, Musk has described a system requiring about 10 million tons of AI satellites to be launched into orbit each year to produce 1 terawatt of computing power. Each satellite has been described as generating about 100 kilowatts.

The proposed orbital systems are intended to rely heavily on solar energy. Musk has argued that suitable orbital configurations could provide greater access to sunlight and reduce the need for battery storage compared with terrestrial data centers. The concept depends on the cost and scale of future launches, satellite production and orbital infrastructure.

What Could Limit Terafab

Building a semiconductor complex at this scale requires specialized equipment, substantial energy, large amounts of land and workers with advanced semiconductor manufacturing skills. Tesla and SpaceX have experience in vehicles, rockets and large industrial facilities, but semiconductor fabrication requires different manufacturing capabilities.

Intel joined the Terafab project in 2026, bringing semiconductor manufacturing experience to an initiative whose original participants had limited experience operating advanced chip fabrication facilities. Tesla also began recruiting engineers for semiconductor process, integration and silicon roles after the project was announced.

The project therefore depends on more than constructing a large building. It requires chip design, fabrication equipment, packaging, testing, energy infrastructure and a workforce capable of operating these systems at scale.

Terafab is best understood as a long-term infrastructure project intended to secure large quantities of computing hardware for robotics, autonomous systems and space-based computing. Its proposed scale, energy requirements and technical demands make execution a substantial part of the challenge, so the stated production targets should be treated as project goals rather than completed results.

The practical point is that Terafab is intended to give Musk’s companies greater control over the production of specialized computing hardware while supporting projected demand from robots, vehicles and orbital data centers. The clearest developments to watch are construction progress, semiconductor manufacturing capability, power infrastructure and the transition from planned capacity to actual chip production.