SpaceX agrees to a ~$40B equipment finance lease with Apollo to fund purchases of Nvidia GPUs for its AI initiatives, set to close before the end of 2026
Carmen Arroyo /Bloomberg:
Context & Ripple Effects
SpaceX had been pursuing a roughly $40 billion chip-financing package earlier in October, split between bank loans and investment-grade debt. The Apollo lease converts that planned GPU funding effort into an equipment-finance structure rather than a conventional balance-sheet purchase.
The financing sits alongside SpaceX's expanding role as a compute supplier: it had agreed to provide Reflection AI access to Nvidia GB300 systems in a compute contract worth up to $6.3 billion, while Anthropic disclosed substantial agreements for SpaceX-hosted Nvidia compute through 2029.
First-order effects
- SpaceX obtains a dedicated route to acquire Nvidia GPUs for its AI initiatives, with Apollo financing the equipment and taking lease exposure once the transaction closes.
- Apollo becomes the capital provider behind a large GPU fleet, while Nvidia gains a financing-backed buyer for additional AI hardware.
Second-order effects
- SpaceX's compute business gains a way to add capacity without funding the full hardware purchase upfront, supporting its ability to serve customers with multiyear Nvidia-compute agreements.
- The structure gives other large AI infrastructure buyers a concrete alternative to unsecured borrowing or equity funding: pair GPU purchases with an equipment lease backed by expected compute demand.
Third-order effects
- If this model is repeated, GPUs increasingly become financeable infrastructure assets, shifting part of AI expansion from corporate capital spending toward specialist-credit underwriting.
- That would make access to debt capacity, customer contracts and hardware residual-value assumptions as consequential to AI buildouts as access to the chips themselves.
The trend: AI infrastructure is moving toward project-style equipment finance, with contracted compute demand used to support increasingly large GPU acquisitions.