SpaceX has secured a binding commitment for a roughly $40 billion master finance lease with an entity managed by Apollo Global Management, the largest committed GPU financing on record. Apollo will buy Nvidia's next-generation Vera Rubin chips and related server hardware and lease them to SpaceX, with each lease schedule running about five to five and a quarter years from delivery. The deal is expected to close before year-end, and leasing begins in the first half of 2027 as equipment arrives (Reuters, 2026).
The financing syndicate includes Goldman Sachs as sole structuring agent, with anchor commitments from BlackRock, Goldman Sachs Asset Management, and KKR. SpaceX will pay rent equal to 5 percent of the financed cost of each tranche plus a fixed annual rate on the outstanding balance. The structure converts what would be a capital expenditure burden into an operating cost, diversifying SpaceX's funding sources as it scales compute for Grok and its cloud services.
SpaceX's AI segment generated $7.6 billion of the company's $14.6 billion in third-quarter revenue, more than Starlink and rocket launches combined. Capital expenditures reached nearly $20 billion, and CFO Bret Johnsen said the company expects to spend $35 billion to $40 billion in the fourth quarter alone. Johnsen cited less than one-year payback periods on AI infrastructure investments to justify the pace.
Musk used the announcement to draw a contrast with pure-play neoclouds, pointing to SpaceX's investment-grade status and cash flow from Starlink and launches as advantages that debt markets have not extended to AI infrastructure companies without major revenue streams.
BuiltWorld AI Operational Take: The lease structure shifts ownership risk of depreciating chips onto Apollo while SpaceX retains use, a sensible move given GPU obsolescence cycles that run shorter than the five-year lease term. The operational bet is that Vera Rubin delivery lands in the first half of 2027 as scheduled. Any delay in Nvidia's production ramp pushes lease commencement and the payback clock, and SpaceX's sub-one-year return assumption depends on putting those chips to work immediately.
