SpaceX targets in-house GPUs as it warns investors of chip supply, costs
Ahead of SpaceX’s expected $1.75 trillion IPO this summer, the company has warned potential investors about its big spending plans to develop artificial intelligence and other technologies.
It lists “manufacturing our own GPUs” among the “significant capital expenditures” it has undertaken, according to excerpts from its S-1 filing reviewed by Reuters. Companies file this document with the U.S. Securities and Exchange Commission to disclose their risks and financials before going public.
Also Read: SpaceX’s IPO hasn’t happened yet, but here’s how investors can make money from Space right now
SpaceX did not immediately respond to a request for comment and the size of the expected spending could not be determined.
The ambition follows efforts by SpaceX, its xAI unit and Tesla to jointly develop Terafab, an advanced AI chip manufacturing complex that CEO Elon Musk is planning in Austin, Texas.
Although Musk said the project will target chips for cars, humanoid robots, and space-based data centers, many details are unknown, including the types of AI chips such as GPUs it will produce. There are various approaches for chips that power artificial intelligence. For example, Nvidia makes GPUs that are largely general-purpose and good at handling a wide range of data processing tasks. Alphabet’s Google is taking another approach, with tensor processing units (TPUs) tuned to perform specific functions, key to building AI models and powering chatbots like Anthropic’s Claude.
It was unclear when SpaceX plans to produce its own chip and which companies (Terafab developers or partners Intel) will manage the manufacturing technologies at the facility.
Musk told Tesla analysts on Wednesday that by the time Terafab scales up, Intel’s next-generation 14A manufacturing process “will probably be pretty mature or ready for prime time” and “seems like the right move.”
It was also unclear whether SpaceX used the term GPU in its filing as an abbreviation for artificial intelligence processors in general.
Still, the previously unreported plans for GPU production come after SpaceX warned investors it may not have enough chip supply to grow.
SUPPLY CONCERNS
“We do not have long-term contracts with many of our direct chip suppliers,” SpaceX said in its S-1 filing. “We expect to continue to source a significant portion of our computer hardware from third-party suppliers and there can be no assurance that we will be able to achieve our TERAFAB-related objectives within the expected time frames or at all.”
Producing GPUs is not easy. Industry heavyweight Nvidia pioneered GPU design and, like much of the industry, outsources their manufacturing to Taiwan’s TSMC.
Also Read: Musk takes home $54,080 in salary in 2025, while Chairman Gwynne Shotwell earns $85.8 million from SpaceX
TSMC has spent billions of dollars and years developing state-of-the-art manufacturing processes for cutting-edge chips that require exotic materials and execute more than a thousand steps with atomic precision. Apple’s years of producing billions of iPhone chips have given it much of the hands-on experience needed to produce cutting-edge processors.
The chip industry, as it is organized, now splits steps such as manufacturing, packaging, and testing among several separate companies. Musk said Terafab will handle every stage of chip production, including design.
