SpaceX Wanted 30,000 Nvidia Racks but May Get 20,000: Why Is Musk Warning Rival AI Cloud Firms?

SpaceX has agreed a binding commitment for an approximately $40 billion (£30 billion) master finance lease for GPUs and related equipment with an entity managed by Apollo Global Management, the company announced on 26 October. If fully drawn, it would be one of the largest known financing packages for dedicated artificial intelligence computing infrastructure.

SpaceX said the arrangement would support the expansion of its terrestrial computing infrastructure, including GPU deployments for Grok and for existing and future cloud services customers.

Elon Musk attributed SpaceX's ability to secure the financing to its investment-grade standing and revenue from Starlink and launch services. He warned that more than 200 neoclouds, the specialist companies that rent out AI computing power, could face 'guppy-level death and consolidation'.

SpaceX's Expected Share of Nvidia's New Racks

Separately, Jeff Pu of GF Securities estimates that SpaceX could receive about 20,000 of Nvidia's Vera Rubin NVL72 racks in 2027, out of projected global supply of 96,000. Pu reportedly said SpaceX had originally sought 30,000.

These figures are analyst estimates. Neither SpaceX nor Nvidia has confirmed them as orders, allocations or delivery schedules, and no delivery contracts have been disclosed.

Nvidia describes the NVL72 as a rack-scale system combining 72 Rubin GPUs with Vera CPUs, networking, power and liquid-cooling components. The systems are designed to be installed as complete units in data centres.

If the estimates prove accurate, 20,000 racks would represent about 21 per cent of the forecast 2027 total. Pu also estimates that six leading neocloud providers will buy a combined 34,000 racks, about 35 per cent of projected supply. SpaceX's reported allocation would equal roughly 59 per cent of that figure. It is unclear whether the six companies include SpaceX, so direct comparisons with individual neocloud allocations remain uncertain.

The financing has not been publicly linked to the reported rack allocation. Nor has a confirmed price for the systems been disclosed, so the amount SpaceX might spend on that hardware cannot be calculated from the available information. Without official pricing or order details, it is not possible to determine how much of the 40 billion US dollars facility could be tied to Nvidia equipment rather than other infrastructure.

Musk's Warning to Neocloud Rivals

Musk argued that neocloud providers without substantial revenue from other businesses could find financing harder to secure. Reuters reported that the warning hit neocloud stocks and heightened concerns about debt among investors following his comments.

The agency said companies including CoreWeave and Nebius had responded by highlighting differences in their business models and funding arrangements. Individual exposure to borrowing costs and the timing of customer revenue will vary according to each company's contracts and financing structures.

If borrowing becomes more expensive and access to high-end AI hardware remains constrained, neocloud providers could face added pressure on their expansion plans. The availability of long-term financing and the ability to sign customers on terms that match repayment schedules are likely to be key factors for smaller operators.

Whether SpaceX's financing translates into a lasting advantage will depend on hardware availability, deployment schedules, financing costs and demand for its services. The company has not detailed how much capacity will be reserved for Grok, how much might be offered to outside customers or how its plans compare with established hyperscale cloud providers.

SpaceX has not set out how its computing infrastructure will relate to Starlink's satellite network, and the announcement does not describe an integrated service. The company has not provided technical or commercial timelines for when the financed equipment will enter service.