Elon Musk announced on X that SpaceX has secured $40 billion in debt financing to acquire graphics processing units that power artificial-intelligence workloads. The funding is intended to purchase GPUs, a critical component for the company's expanding AI initiatives. Musk emphasized that the deal reflects SpaceX's confidence in its long-term revenue streams and its ability to meet the capital demands of next-generation AI models.
According to Musk's follow-up post, the financing consists mainly of callable debt slated to be disbursed during the first two quarters of 2027. A special-purpose vehicle will hold the obligations, keeping them off SpaceX's balance sheet and preserving its investment-grade credit rating. The structure is designed to provide flexibility for future capital needs while avoiding dilution of equity.
Bloomberg, citing sources familiar with the matter, reported that SpaceX has been negotiating with banks and investors to raise the $40 billion, with Goldman Sachs leading the discussions and Apollo Global Management among the potential financiers. The talks focus on a financing arrangement that would fund the purchase of Nvidia GPUs.
The deal highlights the massive capital requirements of the current AI boom, as technology firms pour billions into advanced chips and data-center infrastructure to support increasingly powerful models. GPUs such as those from Nvidia have become essential for training and inference, driving demand that rivals traditional cloud-computing spending.
SpaceX has positioned its AI division, dubbed “Super Intelligence,” as its largest business unit this year. Third-quarter revenue from the segment jumped ninefold to $7.56 billion, up from $813 million a year earlier, according to the company's latest earnings release. The surge reflects growing demand for satellite-enabled connectivity and launch services that support data-intensive AI workloads.
Musk warned that debt markets have largely withdrawn financing from firms lacking substantial revenue streams outside AI data centers. He forecast that more than 200 so-called “neocloud” providers, which specialize in AI-focused cloud infrastructure, could face widespread failures and consolidation, leaving only a few survivors.