SpaceX has entered into a binding commitment for roughly $40 billion in equipment financing to fund advanced artificial intelligence chips and data center infrastructure, according to a regulatory filing released Monday.
Under the transaction, an entity managed by New York-based private equity firm Apollo Global Management will purchase the high-end server hardware and lease it back to SpaceX. In an official disclosure covered by TechCrunch's initial reporting, SpaceX described the transaction as the largest committed graphics processing unit (GPU) financing agreement executed to date.
Structure of the Apollo Equipment Lease
The financing model allows SpaceX to secure specialized computing capacity without absorbing the immediate capital drain of buying hundreds of thousands of state-of-the-art processors outright on standard corporate credit lines.
According to regulatory disclosures, the leases will run for approximately 5 to 5.25 years, with the terms activating sequentially as individual tranches of server hardware arrive. Subject to customary closing conditions, SpaceX expects the master agreement to formally close by the end of the year, with hardware deliveries and lease payments beginning in the first half of 2027.
Apollo confirmed that the capital commitment will directly supply Nvidia’s next-generation Vera Rubin computing platforms. The private equity manager noted that these processors are earmarked specifically for SpaceX’s "terrestrial compute buildout" to expand the physical server farms powering its artificial intelligence models and enterprise cloud offerings.
"This lease diversifies our funding sources and addresses a notable portion of the funding required to deliver on our growth plans," SpaceX Chief Financial Officer Bret Johnsen said in a statement accompanying the filing. Johnsen added that SpaceX has realized "less than one-year payback periods" on its AI compute investments as demand for its infrastructure accelerates.
| Transaction Parameter | Verified Detail |
|---|---|
| Total Financing Value | Approximately $40 billion |
| Financing Partner | Apollo Global Management |
| Hardware Procured | Nvidia next-generation Vera Rubin chips & data center racks |
| Lease Duration | ~5 to 5.25 years per equipment tranche |
| Target Deal Close | By end of 2026 (subject to closing conditions) |
| Operational Deployment | First half of 2027 |
| Target Infrastructure | Terrestrial data centers supporting Grok and cloud services |
Shifting Capital Focus Toward AI Compute
The multi-billion-dollar lease highlights a structural transformation in SpaceX's top-line revenue distribution. While the company built its reputation on orbital launch services and satellite-based internet connectivity through Starlink, artificial intelligence workloads now generate the majority of its business.
During the third quarter, SpaceX’s dedicated AI segment—which includes the Grok chatbot ecosystem, the Cursor developer coding platform, and third-party cloud compute clusters—generated $7.6 billion out of the company’s total $14.6 billion in quarterly revenue. For the first time, AI services accounted for more revenue than the commercial rocket launch manifest and consumer Starlink subscriptions combined.
To keep pace with computing demands, capital spending has expanded in parallel. Quarterly capital expenditures jumped to nearly $20 billion during the third quarter, up from $6.9 billion reported in the identical period the prior year. During an investor earnings call on Oct. 20, Johnsen projected that fourth-quarter capital expenditures would range between $35 billion and $40 billion as compute installations accelerate.
By routing the purchase of physical compute racks through an institutional partner like Apollo, SpaceX preserves capital flexibility while securing guaranteed delivery schedules for silicon architectures currently facing industry-wide supply bottlenecks.
Market Impact and Industry Outlook
SpaceX's agreement sets a legal and financial precedent for how hyper-scale AI operators construct capital assets. As leading model developers encounter steep hardware upgrade cycles every 18 to 24 months, outright ownership of physical silicon poses balance-sheet depreciation challenges. Moving hardware into multi-year leasing structures transfers asset-holding risks while maintaining guaranteed operational run-time.
Apollo's participation also underscores the entry of alternative asset managers into direct technology infrastructure funding, a domain historically dominated by commercial bank consortiums and venture debt syndicates. With Apollo acquiring and holding the underlying silicon assets, institutional investors gain recurring, asset-backed lease distributions secured by SpaceX's broader cash-flow generation.
Hardware integration will proceed over the next several quarters, with the initial Nvidia Vera Rubin racks slated for installation across SpaceX's terrestrial server hubs during early 2027.
