
Elon Musk has unveiled ambitious plans for a $20 billion semiconductor manufacturing plant, named "TeraFab", to produce chips for a wide range of futuristic applications, including space-based data centers, advanced Tesla vehicles, and humanoid robots. The facility will be located in Austin, Texas, next to Tesla’s headquarters, and will operate as a joint venture between Tesla and SpaceX.
In a live-streamed announcement on his social media platform, X, Musk described the project as "the most epic chip-building exercise in history by far." He emphasized the necessity of the facility, explaining, "We either build the TeraFab, or we don’t have the chips, and we need the chips, so we build the TeraFab."
The TeraFab factory is Musk’s solution to the increasing demand for semiconductors at both SpaceX and Tesla. In the broadcast, he outlined the computational requirements of his companies, stating they would need up to 200GW of compute power annually for Earth-based operations and as much as 1TW for space-based processing. Musk also remarked on the global AI compute power available today, saying, "There is currently 20GW of AI compute power available globally", though some estimates place that figure closer to 30GW.
According to Musk, the factory will produce two types of chips: one designed for Edge inference workloads, suitable for vehicles and robots, and a more powerful line of semiconductors hardened for use in the extreme conditions of space.

The space-grade chips produced by the TeraFab will play a crucial role in SpaceX’s ambitious plans for a constellation of orbital data centers. Earlier this year, SpaceX filed for regulatory approval to deploy up to one million data center satellites. Musk previously claimed that this satellite network could provide hundreds of gigawatts of compute power, reinforcing the need for a facility like TeraFab.
The factory is set to be fully integrated, handling all stages of semiconductor production, including logic and memory chip fabrication, packaging, and testing.
While the announcement has garnered significant attention, it also raises questions about feasibility, given Musk’s lack of experience in semiconductor manufacturing and his track record of making ambitious promises that are not always realized. Notably, no specific timeline for when the TeraFab would become operational was provided.
This is not Musk’s first foray into chip design. Tesla previously developed the Dojo D1 chip for processing vehicle data and training its autonomous systems. However, the Dojo project was discontinued in August 2025, with Tesla opting to use chips from established vendors like Nvidia and Samsung instead. Additionally, last year, Musk’s AI venture, xAI, was reported to be hiring staff to develop custom silicon.
Musk’s bold declaration signals a potentially transformative step in addressing the growing demand for chips in advanced technologies, though the success of the TeraFab will ultimately depend on overcoming the challenges of the semiconductor industry’s complex and capital-intensive nature.



