Musk Says SpaceX Closed $40 Billion GPU Debt for 2027

By Rebecca Stone

SpaceX has closed $40 billion in debt financing for graphics processing units (GPUs) expected to arrive in the first half of 2027, Elon Musk said on October 25.

Musk disclosed the financing on X while responding to investor Chamath Palihapitiya, who had warned that mounting financing pressure and delays in data-center construction could leave AI computing capacity idle next year. Musk said he agreed with that risk.

Musk broadly agreed and said SpaceX was better positioned than many competitors. He argued that debt markets had largely shut off funding for companies without substantial revenue independent of AI data centers. SpaceX had an advantage, he said, because it held investment-grade status and generated revenue from Starlink satellite internet and its space businesses.

“That means the 200+ neoclouds are going to see guppy-level death and consolidation. Some will make it through, but not many,” Musk wrote.

What does “closed” mean?

Musk’s post leaves one commercial term unresolved: what “closed” meant. Musk did not say whether investors had agreed to the final debt pricing, and he did not identify the specific GPU models or quantities covered. The $40 billion therefore stands as his statement that financing had closed, not as a disclosed bill of materials for confirmed hardware.

SpaceX CFO Bret Johnsen said during the company’s Q3 2026 earnings call that it expects $35 billion to $40 billion of capital expenditure in Q4 because additional capacity was coming online that quarter. That range is a company forecast, not confirmed spending.

Musk separately narrowed the timing: “This is primarily callable debt for GPU purchases that will arrive in Q1 and Q2. Not a major factor this quarter.” The bulk of the borrowing therefore targets first-half 2027 deliveries rather than the fourth-quarter spending cycle.

How much capacity could the debt fund?

Earlier in October, SpaceX reportedly proposed a $40 billion debt package for Nvidia hardware. The estimate attached to that proposal said the package could cover roughly 360,000 Rubin GPUs across 5,000 Nvidia NVL72 racks, but it did not turn that estimate into confirmed orders.

The reported financing mix consisted of:

• $10 billion in bank loans.
• $30 billion in investment-grade debt.

Musk now says the financing has closed, but his disclosure did not confirm that the reported mix, Rubin count or NVL72 rack count remains unchanged. Those figures describe the earlier proposal and its estimated reach, not verified purchases.

The GPUs fit a broader expansion program, although the capacity claims mix plans with reported deployments:

• In August, xAI outlined plans to increase data-center capacity sevenfold by 2027, targeting 10 gigawatts.
• In September, reporting said SpaceXAI planned to add another 660,000 Nvidia GPUs during 2026, potentially bringing its total to around 1.44 million if the planned deployments proceed.

GPUs need power, cooling, networking equipment and physical infrastructure before they can do useful work. That dependency gives Palihapitiya’s construction-delay warning direct commercial force: financed chips without facilities or customers would increase costs without adding billable computing capacity.

Can SpaceX turn capacity into revenue?

SpaceX has begun selling access to its own infrastructure. In May, Anthropic secured access to the Colossus 1 supercomputer, which includes more than 220,000 Nvidia GPUs and over 300 megawatts of computing capacity.

Later in June, SpaceX reached a separate agreement with Google under which Google would pay $920 million per month for access to approximately 110,000 Nvidia GPUs and related hardware. The payment schedule runs from October 2026 through June 2029.

Unlike the capacity roadmaps, the Google agreement specifies both the monthly price and the payment window. The commercial sequence now demands that SpaceX deliver financed GPUs in Q1 and Q2 2027, bring enough power and data-center capacity online to make them useful, and fill that capacity with paid workloads.

On-time execution could accelerate consolidation among the 200-plus neoclouds Musk expects to face “guppy-level death”; delays would leave SpaceX carrying debt and infrastructure costs against underused capacity.