Elon Musk’s AI ambitions are converging on multiple fronts simultaneously. SpaceX is considering spending up to $119 billion on a semiconductor facility in Grimes County, Texas, dubbed “Terafab” — a vertically integrated chip manufacturing complex developed alongside Tesla and Intel. The facility is intended to produce chips for AI servers, satellites, autonomous vehicles, and SpaceX’s proposed orbital data centres. Musk has argued the factory is an existential necessity, stating his companies cannot secure sufficient chips from existing manufacturers to meet their AI and robotics demands. Grimes County remains one of several locations under consideration.
The compute strategy extends beyond manufacturing. In a surprise move, xAI struck a major partnership with Anthropic, selling the Claude-maker all available capacity at its Colossus 1 data centre — roughly 300 megawatts — after xAI migrated its own training workloads to the newer Colossus 2 facility. The deal, likely worth billions, immediately allowed Anthropic to raise its usage limits and repositions xAI less as a rival AI model developer and more as a neocloud provider renting compute to competitors. Unlike Google, which recently acknowledged constraining Cloud revenue to prioritise its own AI development, and Meta, which built an entirely new cloud infrastructure to fuel Mark Zuckerberg’s AI ambitions, xAI appears willing to monetise spare capacity rather than reserve it for proprietary model development.
That strategic shift comes as Musk’s personal history with AI faces courtroom scrutiny. OpenAI president Greg Brockman testified for two days in Musk’s ongoing lawsuit against his former co-founders, recounting a 2017 confrontation in which Musk demanded unequivocal control of OpenAI’s planned for-profit subsidiary. When the other founders refused, Musk abruptly withdrew his support and eventually left the board in early 2018. Brockman drew on personal journal entries to describe the rupture, which ultimately led OpenAI to raise $1 billion from Microsoft in 2019 — the foundation of its rise as the world’s leading AI frontier lab, and the source of Musk’s enduring grievance.