SpaceX Agrees to Pay $40 Billion to Lease Chips from Apollo

The cost of acquiring advanced AI chips drops significantly, making large-scale AI infrastructure more accessible.

SpaceX has entered into a binding commitment for an approximately $40 billion master finance lease of GPUs and related equipment with an entity managed by Apollo Global Management. This agreement, the largest GPU financing deal ever, will secure next-generation NVIDIA Vera Rubin chips for SpaceX's terrestrial compute infrastructure. The financing is led by Apollo and Goldman Sachs, with anchor commitments from BlackRock, Goldman Sachs Asset Management, and KKR. SpaceX anticipates a payback period of less than one year on these investments, aiming to deploy the world's largest compute infrastructure to support its AI initiatives, including Grok for enterprises and consumers, alongside existing and future cloud services. The lease is expected to close before the end of 2026, with equipment deliveries beginning in the first half of 2027.

This massive financing deal signals a significant acceleration in the demand for and investment in AI compute infrastructure, particularly for specialized chips like NVIDIA's Vera Rubin. Apollo's substantial commitment, alongside other major financial players, highlights the financial industry's growing confidence in the economics of AI hardware and the long-term viability of large-scale AI deployments. The rapid payback periods cited by SpaceX suggest a strong and immediate market for advanced AI services, driven by both consumer and enterprise adoption.

The transaction could pressure other cloud providers and AI companies to secure similar financing or accelerate their own hardware procurement strategies to remain competitive. It also points to a potential shift in how major compute infrastructure is financed, moving beyond traditional capital expenditure models towards large-scale, specialized leasing agreements. This could impact the competitive landscape for AI services and the supply chain for high-end GPUs.