Goldman Sachs and Apollo are assembling a $40 billion financing package tied to SpaceX artificial intelligence chip ambitions, a deal size that places private credit squarely at the center of the global compute race rather than at its periphery, and it signals how quickly infrastructure lending has been reshaped by machine learning demand

Goldman Sachs and Apollo are working on a $40 billion financing structure connected to SpaceX's artificial intelligence chip plans. The scale is the story, because capital of that magnitude rarely moves for speculative projects. Private credit has become a primary funder of compute infrastructure. Banks that once led such deals now share the table. Chip supply remains constrained, which means financing terms often include delivery guarantees.

AI hardware consumes enormous amounts of money before it produces measurable returns. Chips, cooling, land and power contracts all require commitments years ahead of revenue. Lenders accepting that timeline are effectively underwriting a technology bet rather than a cash flow. Structures must therefore blend equity, debt and staged disbursements. Power contracts are becoming as important as silicon in these capital stacks.

The macroeconomic implication is that compute is becoming a capital markets asset class. That brings discipline and disclosure, along with pressure to show returns quickly. Whether the buildout earns its cost will depend on demand holding up long enough to amortize the hardware. Borrowers will be judged on execution speed, because hardware loses value while it waits in a warehouse.