AivexaNewsSearch
AI news for builders and product teamsChecked every hour

SpaceX locks in $40B GPU financing deal with Apollo

Collected Oct 26, 2026

SpaceX disclosed in a regulatory filing on Monday that it has a binding commitment for roughly $40 billion in financing for AI chips and related data center equipment. Under the structure, an entity managed by Apollo Global Management will purchase the GPUs and associated server hardware and then lease the equipment back to SpaceX. SpaceX described it in a statement as the largest committed GPU financing. The company expects the transaction to close by year-end, subject to closing conditions.

The leases run about five to five-and-a-quarter years, with each term starting when a given tranche of equipment is delivered. Leasing is slated to begin in the first half of 2027 as the GPUs and related gear arrive. Apollo said the financing will help deliver Nvidia's next-generation Vera Rubin chips for SpaceX's terrestrial compute buildout.

This is effectively a sale-and-leaseback applied to hardware rather than real estate, and the structure is the interesting part. Instead of borrowing against its balance sheet, SpaceX gets a partner to own the equipment and pays over time for the use of it. That keeps the obligation off the traditional debt ledger in form, though in substance it is a multi-year commitment to make payments, so developers and finance watchers should read it as a long-dated, fixed-cost claim on future cash flow. The tranche-by-tranche lease timing in 2027 also means the compute arrives well after the financing closes, which matters for anyone tracking how quickly SpaceX can convert this capital into usable capacity.

SpaceX CFO Bret Johnsen framed the arrangement as a way to broaden the company's funding base and cover a meaningful share of the capital needed for growth plans, and he pointed to payback periods of less than one year on its AI infrastructure investments as it scales compute for Grok and its cloud services. The company is also spending heavily alongside its Starlink and launch businesses, which the lease is meant to help fund without leaning only on conventional debt markets.

That spending is now large. SpaceX's AI segment, which includes the Grok chatbot, the Cursor coding tool and its cloud compute business, brought in $7.6 billion of the company's $14.6 billion in third-quarter revenue, more than Starlink and rocket launches combined, and outperformed investor expectations. Capital expenditures climbed to nearly $20 billion from $6.9 billion a year earlier. On the October 20 earnings call, Johnsen said SpaceX expects to spend $35 billion to $40 billion in the fourth quarter alone.

The backdrop is a shakier market for AI infrastructure debt. The article notes that investors have grown uneasy about steep capital expenditures and thin profits among AI infrastructure players. Elon Musk used the announcement to argue that SpaceX's investment-grade status and cash flow from Starlink and launches give it access that pure-play compute companies lack, saying on X that more than 200 neoclouds, businesses built solely to rent out GPUs for AI workloads, face what he called guppy-level death and consolidation, with some surviving but not many.

It is worth separating the reported numbers from the rhetoric. The $40 billion figure, the lease length, the 2027 start and the Vera Rubin reference all come from the filing and statements. Musk's prediction about neocloud consolidation is a forecast from an interested party; as a reading of the market, it is plausible that rising capital costs squeeze operators who lack non-AI revenue, but the source gives no data on individual neocloud finances. The comparison Musk draws is also somewhat apples-to-oranges: SpaceX is financing its own compute for Grok and cloud services, while neoclouds finance compute to rent to third parties, so their risk profiles differ even if both depend on cheap capital.

A likely trade-off for SpaceX is flexibility versus cost. Leasing avoids a large upfront outlay and diversifies funding, but multi-year lease payments become a recurring obligation that must be covered even if AI demand softens. Naming Nvidia's Vera Rubin generation ties the deal to hardware that is not yet broadly deployed, so delivery timing becomes a schedule risk. For product teams, the practical read is that SpaceX is signaling sustained compute expansion for Grok, Cursor and its cloud business into 2027 and beyond.

Why it matters: The deal shows a large AI operator funding chips through leasing rather than straight debt, which could become a template for capital-hungry infrastructure buyers as lenders get pickier. Developers building on Grok, Cursor or SpaceX cloud capacity should expect continued compute buildout, though the leases beginning in 2027 suggest capacity arrives on a multi-year schedule. If Musk's read on neocloud consolidation proves right, teams sourcing GPU capacity from smaller rental providers may face fewer vendors and changing prices, an inference rather than a stated outcome.

Read at TechCrunch · AI

Based on reporting from the original publisher. Visit the source for full context and later updates.

Publisher excerpt

The financing gives SpaceX another way to fund its rapidly expanding AI ambitions without relying solely on traditional debt markets.