SpaceX plans to release its own graphics processor in partnership with Tesla, but considers the project risky.

SpaceX is preparing for an IPO and discloses risks to regulators

When preparing a public offering, SpaceX must provide U.S. authorities with a report on potential threats that could affect its business. In this document (S‑1), it is emphasized that the project to develop proprietary graphics processors (GPU) is considered one of the most risky.

Plans for GPU Release

SpaceX intends to jointly develop and manufacture graphics chips with Tesla to support AI infrastructure. This will be carried out at a new Terafab plant planned for construction in the coming years. The report states that launching production will require significant capital investments.

Technical Aspect

Elon Musk has not previously specified which exact chips will be produced at Terafab. Earlier this week he announced that Tesla would use Intel’s 14 A technology for its own processors, but the location of their manufacturing remains unknown. Unlike the AI‑5 on‑board chips developed solely for Tesla vehicles, “GPU” implies a more complex architecture.

Musk described Terafab as a comprehensive facility: not only wafer processing but also packaging finished chips with subsequent testing, including memory chip integration. Nevertheless, the term “GPU” can apply to any AI chip, so it is premature to draw conclusions about Tesla’s specific lineup.

Risks and Dependencies

SpaceX currently has no long‑term contracts with chip suppliers and will be heavily dependent on external manufacturers. This adds significant risk for the company. The report stresses that there is no guarantee the Terafab project will finish on time or in the required volume.

Similarly, TSMC has already noted that Tesla’s path to mass production of its own chips will be difficult and slow. This confirms that launching in‑house semiconductor manufacturing remains a serious challenge for both companies.