SpaceX Commits $40 Billion to Nvidia GPUs Backed by Starlink and Launch Revenue

SpaceX just put a number on the cost of turning a rocket-and-satellite company into one of the world’s biggest artificial-intelligence operators: roughly $40 billion.

The company announced Monday that it has reached a binding commitment for a master finance lease covering Nvidia GPUs and related equipment.

A $40 billion lease sits far beyond a routine server order. SpaceX calls it the largest committed GPU financing ever.

The more revealing part is how Elon Musk says SpaceX got through a debt market that is closing around smaller AI-compute companies. Starlink subscriptions and launch revenue gave lenders something beyond a promise about future demand.

Musk made that distinction bluntly:

Pretty much.

As of the past few weeks, debt markets have almost entirely shut off access to money for any company that doesn’t have substantial revenue independent of AI data centers.

SpaceX closed $40B of debt for GPUs, but we’re investment grade and have Starlink & launch revenue. That means…

— Elon Musk (@elonmusk) October 25, 2026

SpaceX says the lease will deliver Nvidia’s next-generation Vera Rubin chips for a terrestrial compute buildout. The stated goal is to add computing capacity for Grok, enterprise products, consumer services and future cloud customers.

The agreement is a binding commitment, but it is not yet a completed closing. SpaceX says specified closing conditions remain and expects the master lease to close before the end of 2026.

Individual lease schedules would begin as GPU pods and related equipment arrive during the first half of 2027. That timeline matters because it separates Monday’s financing commitment from machines that are installed, powered and earning revenue.

SpaceX CFO Bret Johnsen says the company is seeing payback periods of less than one year on its super-intelligence infrastructure investments. That is the company’s claim, and the next test will be whether new demand keeps pace with the enormous hardware bill.

Apollo Global Management confirms that an Apollo-managed entity will sit on the other side of the master finance lease. Apollo led the financing parties, while Goldman Sachs served as sole structuring agent.

Funds and accounts managed by BlackRock, Goldman Sachs Asset Management and capital accounts advised by KKR also made anchor commitments. That roster shows how far the AI infrastructure race has moved beyond ordinary corporate borrowing.

The financiers are betting on more than a chatbot. They are underwriting a company with paying satellite customers, government and commercial launch contracts, reusable rockets and a fast-growing AI operation under the same roof.

The structure also gives SpaceX another way to add hardware without paying the entire purchase price up front. The equipment is financed through lease schedules tied to delivery, pushing the real economic test into 2027 and beyond.

The Associated Press reported last week that SpaceX earned $4.4 billion on $14.6 billion in third-quarter revenue, reversing a year-earlier loss as its satellite communications and AI businesses expanded. AP said AI had become the company’s largest business by revenue, while Starlink remained its biggest profit generator.

That mix helps explain Musk’s point. AI may be consuming the capital, but the launch and connectivity businesses give creditors established operations to evaluate.

It does not make the spending small. AP reported that SpaceX’s third-quarter capital expenditures reached $19.6 billion, nearly triple the prior-year figure, and that the company expects to spend $35 billion to $40 billion in the fourth quarter.

The new lease is therefore not an isolated splash. It is part of an expansion already moving at a scale that would overwhelm most companies.

Musk also clarified when the financed hardware is supposed to matter:

This is primarily callable debt for GPU purchases that will arrive in Q1 and Q2. Not a major factor this quarter.

— Elon Musk (@elonmusk) October 25, 2026

TechCrunch explains that the Apollo-managed entity is expected to buy the chips and related server hardware, then lease that equipment to SpaceX. The leases are expected to run roughly five to five-and-a-quarter years from each delivery tranche.

The outlet also places the deal against a harsher market for so-called neocloud companies that borrow heavily to buy GPUs and rent the capacity to AI developers. Musk predicts many of those pure-play operators will fail or consolidate because lenders increasingly want revenue that does not depend entirely on data-center economics.

His warning is self-serving. SpaceX is entering the same compute market and benefits if weaker competitors lose access to capital.

But the financing commitment makes the competitive advantage concrete. SpaceX can point to broadband subscriptions and launch missions while it asks Wall Street to fund the next wave of AI hardware.

There are still serious questions. The agreement must close, the equipment must arrive, power and data-center capacity must be ready, and SpaceX must convert all that compute into revenue fast enough to justify the lease obligations.

The company also needs to prove Johnsen’s less-than-one-year payback claim at much larger scale. Returns achieved on earlier clusters do not automatically repeat when tens of billions of dollars of new equipment arrive in a compressed window.

Still, Monday’s announcement marks a major shift in what SpaceX is becoming.

The rockets and satellites are no longer just the headline businesses. They are now the financial base helping Musk buy one of the largest piles of AI computing hardware ever assembled.