SpaceX warned of risk of 'non-viability' of AI data centers in space - Reuters

Elon Musk's SpaceX plans to create space data processing centers for artificial intelligence, as well as - the creation of human settlements on the Moon and Mars are based on untested technologies and may not be commercially viable. The company warned of this in a confidential filing with the U.S. Securities and Exchange Commission (SEC) ahead of going public. The document was reviewed by Reuters.

In particular, the prospectus notes that any future orbital data centers for AI will operate "in a harsh and unpredictable space environment, subject to a wide and unique range of space risks that could lead to malfunctions or failures" in operations.

The risk factors, disclosed in the prospectus as required by U.S. securities laws, are designed to inform investors of potential threats while reducing companies' legal liability, Reuters explains.

SpaceX did not respond to the agency's request for comment.

Risks disclosed in previously unreleased SpaceX documents indicate a more restrained assessment of the company's prospects compared to recent public statements by its CEO Elon Musk, the agency notes. This comes as SpaceX prepares for potentially the largest initial public offering in history. SpaceX aims to hold an IPO in June - during it the company expects to raise about $75 billion and get a valuation of about $1.75 trillion.

Musk said at the World Economic Forum in January that creating data centers for AI in space is "the obvious solution." Within two to three years, Musk argued, it would become the cheapest way to put computing power. In February, following the announcement of SpaceX's merger with Musk's AI startup xAI, the businessman said that "space AI is basically the only way to scale" - referring to building up computing power for AI development.

This article was AI-translated and verified by a human editor