SpaceX Seeks $40 Billion Apollo-Led Financing to Buy Nvidia AI Chips, FT Reports

SpaceX is seeking to raise $40 billion in a financing effort led by asset manager Apollo Global Management to fund a blockbuster purchase of Nvidia artificial intelligence chips, the Financial Times reported on Tuesday, citing people familiar with the matter.

According to the FT, the Elon Musk-led company is looking to raise about $10 billion in bank loans and $30 billion in investment-grade debt to pay for the chip order. Apollo is expected to lead the deal and help sell the debt to a broad range of investors, with bond fund Pimco among a small group of lenders in talks to provide financing. The transaction is expected to close in 2027, according to the newspaper’s sources.

SpaceX, Apollo, Nvidia and Pimco did not immediately respond to requests for comment, Reuters reported. Pimco itself declined to comment.

Shares in SpaceX, which trade publicly after the company’s June IPO, fell about 1 per cent in extended trading on Tuesday after the report emerged, and were down almost 2 per cent in premarket trading on Wednesday at around $169.15, according to MarketWatch. Nvidia’s stock edged higher by about 0.5 per cent over the same period, a muted move for the chipmaker despite what would be one of the largest single AI chip orders ever discussed publicly.

The reported fundraising would land among the biggest debt financings ever assembled for the AI buildout, according to Bloomberg, which independently confirmed on Wednesday that SpaceX is in talks with banks and investors for the $40 billion package. Bloomberg’s sources added that the fundraising discussions are in an early stage and could end without a deal being completed.

A borrower almost the size of its own revenue

The sheer scale of the borrowing is striking. MarketWatch, citing FactSet data, noted that the $40 billion would represent essentially all of SpaceX’s projected revenue for this year, around $44.5 billion. For a company that only listed its shares publicly in June in a record $86 billion initial public offering, raising the equivalent of an entire year’s sales in debt to buy chips would be an extraordinary statement of intent — and a signal of just how capital-hungry the AI race has become.

According to Bloomberg, SpaceX’s second-quarter capital expenditure was $18.4 billion, with roughly $15.8 billion — about 86 per cent — going to AI infrastructure, against approximately $1.4 billion for Starlink and $1.2 billion for the space business. The company’s finance chief had previously hinted at full-year capital spending of around $65 billion, Bloomberg reported.

The Nvidia lock-in

The chip order is tied to a strategic bet Musk has already made public. On SpaceX’s August earnings call, the chief executive said the company had “decided to build exclusively on Nvidia” for its AI initiatives, describing the chipmaker’s Vera Rubin platform as the best AI architecture available. The company absorbed Musk’s AI startup xAI in February before going public, and now runs the Grok chatbot and other models through its Colossus data-centre clusters.

Musk said last month that xAI’s Colossus 2 data centre could more than double its Nvidia chip count by December. In a post on X, he disclosed that the facility currently runs 110,000 Nvidia GB200 chips and 440,000 GB300 chips, with another 220,000 GB300 units expected to come online imminently, a further 220,000 in November, and as many as 220,000 more in December “if we get lucky.”

SpaceX also rents out its Colossus computing capacity to outside AI developers, principally Anthropic and Alphabet’s Google, according to market coverage of the financing talks. Super Micro Computer chief executive Charles Liang recently said the company’s data-centre work for SpaceX’s AI operations was moving “full speed ahead.”

Nvidia, for its part, is the dominant supplier of the processors used to train and run large AI models, and demand for its hardware has remained intense as developers race for compute capacity. In August, Nvidia partnered with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR on financing platforms intended to mobilise more than $500 billion for AI infrastructure projects — a move that underscores how closely the chipmaker’s fortunes are now tied to the credit markets that fund its customers.

A wider wave of debt-fuelled AI spending

SpaceX’s planned $40 billion raise is the latest in a rapid-fire series of giant AI infrastructure financings. Bloomberg reported just days earlier that Broadcom’s Wall Street syndicate is starting to gather $60 billion of fresh AI chip financing to benefit Anthropic and other companies. Morgan Stanley estimates that AI infrastructure will require $1.5 trillion in external financing by 2028, even as lenders and investors grow more cautious about funding the industry’s expansion, Reuters reported.

The trend has also drawn warnings. At the Forbes Global CEO Conference in Singapore on Wednesday, Bridgewater Associates founder Ray Dalio told executives that the AI investment boom looks like a “classic bubble” approaching a breaking point, arguing that heavy debt is financing the buildout in a pattern that mirrors past speculative cycles, according to conference coverage. Franklin Templeton’s Jenny Johnson, appearing on the same panel, added that AI has not yet produced real productivity gains for most US companies.

SpaceX’s parallel moves

Separately this week, SpaceX’s AI and satellite operations made other headlines. The Federal Communications Commission granted approval for the company to launch an additional 15,000 next-generation Starlink satellites as part of its Starlink Mobile service, PCMag reported. And on Wednesday morning, Musk announced an update to the Grok Bot AI assistant in a social media post, saying SpaceX would going forward use “the best back end model for any given task” — including rival APIs such as Anthropic’s Claude Opus 5.5, Midjourney and Suno.

Analysis: Why It Matters

The reported $40 billion financing tells three important stories about where the AI boom stands in late 2026 — stories that go well beyond one company’s borrowing plans.

First, it confirms that the AI infrastructure race has moved decisively from equity to debt. A company that raised $86 billion in fresh equity barely four months ago is now looking to borrow an amount almost equal to its annual revenue. That is not the behaviour of a company short on ambition; it is the behaviour of an industry that believes compute demand will outrun supply for years and is willing to mortgage the future to lock in chips today. The structure — $30 billion of investment-grade debt rather than junk bonds — suggests lenders are pricing this as relatively low-risk paper, betting that AI compute capacity will generate reliable cash flows. If Morgan Stanley’s $1.5 trillion financing estimate for 2028 is even close, SpaceX’s deal will be remembered as a milestone in the moment the credit markets became the primary engine of the AI era.

Second, the deal deepens the most consequential vendor lock-in in technology. SpaceX has publicly committed to building exclusively on Nvidia’s architecture, and Nvidia’s August tie-up with the very asset managers now financing SpaceX — Apollo, BlackRock, KKR and others — means the chipmaker is effectively helping fund the financing of its own customers. This is a virtuous circle for Nvidia: it sells the chips, and its finance partners lend its customers the money to buy more. For SpaceX, the exclusivity buys architectural coherence and priority supply in a market where chip availability is the binding constraint. The risk is obvious: if a rival architecture — from AMD, whose chief executive Lisa Su this week predicted “very high” chip demand continuing for years, or from custom silicon players — pulls ahead on price or performance, SpaceX will have borrowed tens of billions against a single vendor’s roadmap.

Third, the market’s reaction hints at a new anxiety. SpaceX’s shares slipped while Nvidia’s rose, suggesting investors are beginning to distinguish between the companies that sell the picks and shovels of the AI gold rush and those that buy them. Dalio’s bubble warning this week may prove premature — warnings of this kind have shadowed every infrastructure buildout from railways to fibre optics, and most of those assets were eventually used — but the direction of travel is hard to ignore: ever-larger amounts of debt, raised against ever-more-expensive chips, to fund capacity whose revenues are still being proved. The space company’s first-half numbers, with AI capex dwarfing all other spending, show a business being remade around the hypothesis that compute will pay.

What to watch next: whether the financing closes in 2027 as reported, and on what terms — any widening of credit spreads would signal lender unease with AI leverage. Whether Nvidia’s Vera Rubin platform delivers the performance SpaceX is betting its balance sheet on. And whether lenders’ appetite for AI debt keeps growing, or whether caution, flagged by Morgan Stanley, starts to ration the boom.

Sources

• Reuters
• Bloomberg
• MarketWatch