Musk claims accelerated chip tool deliveries for Terafab

The semiconductor industry is preparing for a massive shift in infrastructure as Elon Musk claims that key equipment suppliers are fast-tracking deliveries for a new joint manufacturing venture. During a live broadcast of the Q3 2026 earnings call on X on October 20, Musk announced that ASML and other critical partners have expedited shipments to support the Terafab facility.

Terafab is a multi-billion dollar project between Tesla and SpaceX designed to produce one terawatt of AI computing power annually. The primary facility will be located in Gibbons Creek, Texas. Musk stated that the team targets volume chip production there by the third or fourth quarter of 2028. He described this schedule as highly aspirational and noted that it carries significant uncertainty, referring to large error bars in the timeline.

In a direct quote, Musk expressed gratitude for the accelerated support, saying, These are very aggressive dates, and we're very appreciative of ASML and all of our other suppliers in accelerating equipment deliveries in order to achieve this outcome. The project aims to create a vertically integrated factory that handles logic, memory, and advanced packaging in a single location. This internal capability is intended to supply chips for xAI clusters, Tesla’s self-driving vehicles, Optimus robots, and SpaceX’s orbital data centers, reducing reliance on external foundries.

The technical foundation of Terafab relies on Intel’s 14A manufacturing node technology. Intel officially joined the venture in April 2026 to help co-develop silicon fabrication methods. ASML remains the most critical partner as the sole commercial supplier of extreme ultraviolet lithography systems, which cost roughly $400 million each. While unofficial sources indicated the first EUV tool was slated for mid-2027, it is unclear if this date reflects the expedited timeline Musk mentioned.

Beyond ASML and Intel, the supply chain includes American firms like Applied Materials, Lam Research, and KLA Corporation, along with Japan’s Tokyo Electron. These companies provide front-end processing and metrology tools. Tesla and SpaceX are also negotiating with approximately ten South Korean equipment makers, including Hanmi Semiconductor and HPSP, for specialized packaging and piping components.

Musk also outlined broader AI capacity goals, expecting 2.5 to 3 gigawatts of compute to be active by the end of 2026. He expressed high confidence in adding another 10 gigawatts in 2027, aiming for a total of 13 gigawatts. To reach this, SpaceX and xAI are procuring a 440,000-GPU Nvidia Vera Rubin training cluster, which he plans to have online by the end of Q2 2027.

Despite the ambitious plans, industry experts remain skeptical about the one terawatt target. Gartner estimates global AI-optimized data center capacity at roughly 44 gigawatts as of late 2026, meaning Terafab’s annual output would be significantly larger than the entire current global total. The sheer scale of the project highlights the growing demand for domestic chip production in the United States.