SpaceX has finalized a binding agreement for approximately $40 billion in funding dedicated to AI chips and associated data center infrastructure, as the company revealed in a Monday regulatory filing. SpaceX referred to this arrangement as the “largest committed GPU financing” in an official statement.
Under the terms of the agreement, an entity managed by Apollo Global Management, the prominent New York-based private equity firm, will acquire the chips and relevant server hardware, subsequently leasing this equipment to SpaceX. The leases are projected to last roughly five to five-and-a-quarter years, commencing upon the delivery of each equipment tranche, according to the filing.
SpaceX anticipates the deal’s completion by the end of the year, contingent on standard closing conditions, with leases becoming active in the first half of 2027 as the GPUs and related hardware arrive.
This financing mechanism offers SpaceX an alternative method to fund its rapidly growing AI endeavors, reducing its sole reliance on conventional debt markets.
“This lease diversifies our funding sources and addresses a notable portion of the funding required to deliver on our growth plans,” SpaceX CFO Bret Johnsen commented in a Monday statement, further noting that the company is seeing “less than one-year payback periods” on its AI infrastructure investments as it scales computing power for Grok and its cloud services.
Apollo stated that this financing will facilitate the delivery of Nvidia’s upcoming Vera Rubin chips for SpaceX’s “terrestrial compute buildout.”
The funding arrives as SpaceX escalates its expenditures to expand its AI infrastructure, alongside its Starlink and launch operations.
This financing follows a record-setting third quarter for SpaceX’s AI division. SpaceX’s AI segment, which encompasses the Grok chatbot, the Cursor coding tool, and its cloud computing offerings, exceeded investor projections in the third quarter, generating $7.6 billion of SpaceX’s $14.6 billion in total revenue, surpassing the combined revenue from Starlink and rocket launches.
Capital expenditures also saw a significant increase to nearly $20 billion, up from $6.9 billion in the prior year, as SpaceX rapidly expands its computing capacity. CFO Bret Johnsen indicated on the company’s October 20 earnings call that SpaceX projects spending between $35 billion and $40 billion in the fourth quarter alone.
This substantial lease deal comes after a volatile period in the debt markets for companies involved in AI infrastructure, as investors become increasingly wary of escalating capital expenditures and limited profitability.
While SpaceX had no difficulty accessing Wall Street, Musk leveraged the $40 billion announcement to highlight SpaceX’s advantage over specialized AI compute competitors.
Citing SpaceX’s “investment-grade” credit rating and consistent cash flow from Starlink and rocket launches, Musk asserted on Sunday that debt markets have largely become inaccessible to companies lacking significant revenue streams beyond AI data centers—signaling potential challenges for the over 200 neocloud firms established solely to lease GPUs for AI workloads.
“That means the 200+ neoclouds are going to see guppy-level death and consolidation,” Musk posted on X. “Some will make it through, but not many.”