SpaceX closes $40 billion debt deal to buy Nvidia AI chips as Musk warns 200-plus neoclouds could fail or merge

SpaceX has closed $40 billion in debt financing to buy artificial intelligence chips from Nvidia, founder Elon Musk said. The funds will buy graphics processing units from Nvidia, with deliveries due in the first and second quarters of 2027.

How the deal is structured

Musk said SpaceX's investment-grade credit rating, along with revenue from its Starlink satellite business and rocket launches, helped it secure the financing. The debt is mainly callable, meaning it can be repaid early under agreed terms. "Not a major factor this quarter," he wrote on X.

Earlier this month, Reuters cited the Financial Times as reporting that SpaceX was in talks with banks and asset managers about the financing, in a package of about $10 billion in bank loans and $30 billion in investment-grade debt led by Apollo Global Management, with Pimco among the lenders in talks, and a close then expected in 2027. The deal has now closed earlier than that timetable, with deliveries brought forward.

Warning to neoclouds

Musk also warned that debt markets have almost stopped lending to companies without major revenue outside AI data centers. "That means the 200+ neoclouds are going to see guppy-level death and consolidation. Some will make it through, but not many," he wrote on X.

Neocloud firms rent computing power to AI developers and other clients. Many borrow money to buy GPUs and expand their services. Musk expects firms that cannot secure more funding to fail or merge with rivals.

AI spending still accelerating

SpaceX's latest quarterly results showed revenue up 161% year over year to $14.58 billion, with Starlink ending September at 13.6 million consumer subscribers and enterprise and government revenue up 91%. Capital spending nearly tripled to $19.6 billion in the third quarter and is expected to rise to $35 billion to $40 billion in the fourth quarter.

The $40 billion deal captures the AI infrastructure race in miniature: the greater the demand for computing power, the more money goes into buying chips, and the gap between companies that can raise that money and those that cannot is widening.