Musk says SpaceX secured $40 billion for GPUs due in 2027

Elon Musk says SpaceX has closed $40 billion in financing for graphics processing units expected to arrive mainly in the first half of 2027. Reuters reported his statement on 25 October. The announcement points to a large coming hardware programme, while leaving the detailed borrowing terms and drawdown schedule undisclosed.[1]

Musk described the financing as primarily callable debt and said it would not be a major factor in the current quarter. He also credited investment-grade standing and cash generation from Starlink and launch services with supporting the company’s borrowing. Those explanations are his account of the transaction, rather than independent confirmation of its pricing or contractual terms.[1]

Outstanding debt and a later financing announcement

SpaceX's quarterly filing for the period ended 30 September 2026 reported $38.006 billion in aggregate principal amount of debt outstanding at that date. It describes, among other agreements, $25 billion of senior unsecured notes issued in June in several maturity tranches. The new $40 billion statement was made on 25 October—after the financial reporting date.[1][2]

The September figure records debt already outstanding at a balance-sheet date; the October statement describes a subsequent financing arrangement. Commitments, future draws or refinancing could affect how the arrangement appears in later accounts. Adding the two figures would therefore produce a debt total that neither source establishes.[1][2]

From funded chips to working compute

A GPU financing facility is one piece of a longer capital chain. Even if chip purchases are fully funded, building a usable training or inference cluster requires racks, networks, cooling, power contracts, site preparation and operational staff. Chip delivery in the first or second quarter of 2027 is an expected procurement window, not the date every processor becomes productive.[1]

SpaceX’s third-quarter earnings disclosure reported $14.575 billion in revenue. That supplies context for the scale of the business, but it does not measure the return on the newly announced GPU programme. Installation, utilisation and the costs of operating the additional compute will determine how the investment contributes over time.[3]

The terms still to be disclosed

Calling the debt “callable” does not disclose its interest rate, maturities, collateral or the conditions and costs of early repayment. Those terms matter when assessing a programme whose hardware arrives over several months and may become productive later still. The announcement alone cannot resolve them.[1]

There are also at least three risks to separate: procurement risk if the GPU supply chain changes, execution risk if compute sites cannot turn on promptly, and demand risk if purchased compute does not attract enough profitable work. A company can succeed at the first two and still struggle with the third. Conversely, some additional borrowing may finance contracts scheduled against future usage rather than an idle fleet.[1]

What later filings can show

The next useful disclosures would connect borrowing with delivered hardware and operating capacity: how much financing has been drawn, how many processors are installed, and how intensively they are used. Together with depreciation and financing costs, those measures would make the economics of the programme clearer than the headline commitment alone.[1]

Report an error or suggest a correction