The key facts

  • SpaceX has entered a financing agreement worth roughly $40 billion with an Apollo-managed vehicle — chips and servers will be purchased and leased back to the company for 5–5.25 years.
  • The deal is described as the largest committed GPU financing; closing is expected by the end of the year, with leasing starting in the first half of 2027.
  • SpaceX's AI segment brought in $7.6 billion of $14.6 billion in quarterly revenue, while capital expenditures rose to nearly $20 billion.

A $40 billion deal

SpaceX said in a regulatory filing on Monday, that it has reached a commitment agreement for roughly $40 billion in financing to purchase AI chips and related data center equipment. The company called the deal “the largest committed GPU financing”. Under the agreement, a vehicle managed by Apollo Global Management — a New York-based private equity firm — will purchase the chips and server equipment and then lease them to SpaceX for a term of approximately five to five and a quarter years, starting from the delivery of each batch of hardware. SpaceX expects to close the deal by the end of the year, subject to conditions, with lease payments beginning in the first half of 2027 as the GPUs are delivered.

Why SpaceX needs the money

The financing gives SpaceX an additional way to pay for its fast-growing AI projects without relying entirely on traditional debt markets. The company's CFO, Bret Johnsen, said, that the lease “diversifies funding sources and covers a significant portion of the capital needed for our growth plans,” adding that investments in AI infrastructure pay back in less than a year as computing capacity for Grok and cloud services scales up. On Apollo's side, it was noted that the deal will help deliver to SpaceX Nvidia's next-generation Vera Rubin chips for “ground-based compute expansion.”

Context: a record quarter and rising expenses

The agreement followed a very strong third quarter for SpaceX's AI business, which includes the Grok chatbot, the Cursor programming tool, and cloud computing: the segment exceeded investor expectations and brought in $7.6 billion of the company's $14.6 billion in total revenue — more than Starlink and rocket launches combined. Capital expenditures rose to nearly $20 billion, up from $6.9 billion a year earlier, and at the earnings call on October 20 Johnsen said SpaceX expects to spend $35–40 billion in the fourth quarter alone.

Market backdrop and Musk's statement

The large lease deal came after a turbulent period in the debt market for AI infrastructure players: investors are wary of unchecked capital spending and limited profitability. At the same time, SpaceX has had no trouble accessing Wall Street, and Elon Musk used the announcement to celebrate superiority over competitors focused exclusively on AI computing. Pointing to SpaceX's “investment-grade” rating and stable cash flow from Starlink and launches, Musk claimed on Sunday, that debt markets have effectively closed for companies without major revenue sources outside AI data centers — which, according to him, spells trouble for the more than 200 “neoclouds” created solely to rent out GPUs. “This means 200+ neoclouds face minnow-level mortality and consolidation. Some will survive, but few,” he wrote on X.

Why it matters

The deal shows that access to AI infrastructure financing increasingly depends on diversified revenue: SpaceX was able to raise $40 billion thanks to income from Starlink and launches, whereas, in Musk's assessment, narrowly specialized GPU rental operators are losing access to debt. For the industry, it is a signal of possible consolidation among neoclouds, and for SpaceX itself, confirmation of large-scale plans: the company expects to spend $35–40 billion in the fourth quarter alone. In practical terms, this means that deliveries of Nvidia Vera Rubin chips and the start of lease payments are scheduled for the first half of 2027, provided the deal closes by the end of the current year.