SpaceX is arranging a roughly $40 billion financing package led by Apollo Global Management to fund purchases of Nvidia artificial intelligence processors, according to a Yahoo Finance report on the planned transaction.

The reported deal, if completed at the stated size, would rank among the largest single-buyer capital commitments to Nvidia AI silicon disclosed publicly. The Yahoo Finance report did not specify which Nvidia product families — Hopper, Blackwell, or a mix — SpaceX intends to acquire, nor did it provide shipment volumes, wafer-allocation details, or delivery timing.

What did the report actually say?

The Yahoo Finance item identifies Apollo as lead arranger of the financing. It does not name additional lenders, disclose the term length, or describe the structure — whether senior secured, asset-backed against the chips themselves, or a hybrid credit facility. Yahoo Finance did not name a closing date, and SpaceX, Apollo, and Nvidia had not issued public statements as of the report.

Who is Apollo in this deal?

Apollo Global Management operates one of the largest private credit and asset-backed financing businesses in the world, with several hundred billion dollars under management across its credit and yield-focused funds. The firm has been an active financier of digital infrastructure, including data-center campuses, fiber, and structured transactions tied to GPU procurement. A lead-arranger role on a $40 billion AI-chip financing would be unusually large even for Apollo, whose largest disclosed technology financings to date have generally been in the single-digit billions per transaction.

Why would SpaceX need $40 billion in AI silicon?

SpaceX has not historically appeared on Nvidia's published customer lists, which are dominated by hyperscale cloud operators. Through fiscal 2025, Nvidia's data-center segment revenue tracked at a multi-quarter run rate above $100 billion, with Microsoft, Google, Meta, Amazon, and Oracle publicly disclosing the largest commitments to Hopper (H100, H200) and Blackwell (B100, B200, GB200) deployments.

A $40 billion order would meaningfully exceed the annual AI capex of any single hyperscaler, suggesting either a multi-year buildout schedule or a deployment tied to adjacent Musk-affiliated entities. The financing arrives as xAI — Musk's AI venture, separate from SpaceX but sharing its founder — scales its Memphis supercomputer cluster, and as Starlink's ground segment increasingly relies on GPU-accelerated inference.

What isn't yet known?

• Chip mix. Whether the package is concentrated in current-generation Blackwell systems, a forward allocation of next-generation Rubin-family silicon, or a blend.
• Delivery cadence. Whether the $40 billion represents a one-time purchase, a multi-year take-or-pay structure, or an option-style facility that draws against future Nvidia shipments.
• End user. Whether the chips are destined for SpaceX internal workloads, an affiliated entity, or third-party capacity sales.
• Pricing assumptions. Whether the deal is priced at list, or reflects volume discounts typical of large hyperscaler contracts — discounts that can reach 20% or more off Nvidia's published prices.

What changes commercially if the deal closes?

A confirmed $40 billion SpaceX commitment would tighten Nvidia's near-term allocation, particularly for Blackwell GB200 NVL72 rack-scale systems, where supply has been the binding constraint through 2024 and into 2025. It would also signal a new buyer class — vertically integrated operators outside the traditional hyperscaler set — capable of absorbing allocations at scale, with implications for pricing power and product prioritization at Nvidia.

The closing terms, the identity of co-lenders, and the underlying chip SKUs will determine whether this transaction reshapes the AI hardware supply picture or simply locks in capacity already implied by Nvidia's existing backlog.