Elon Musk says SpaceX has closed a $40 billion debt package for GPUs, a deal that would push the rocket-and-satellite company deeper into the AI infrastructure race even as credit investors are already signaling anxiety. The Tesla chief posted on X on Sunday, Oct. 25, 2026, that the financing will help SpaceX buy advanced chips, while also arguing the company can shoulder the load because it has investment-grade status and revenue from Starlink and rocket launches. At the same time, SpaceX credit-default swaps jumped to a record 266 basis points at Friday’s close, a junk-rated level that underscores how much risk the market sees in this stretch of expansion.
The move lands at a moment when SpaceX is growing fast, spending even faster, and making a public case that it can finance AI hardware without losing control of its balance sheet. Bloomberg and MarketWatch reported that the debt is primarily callable for GPU purchases arriving in the first and second quarters of 2027, which means the capital is aimed squarely at a future buildout rather than near-term operations. Musk also said the debt is not a major factor this quarter, a reminder that the company is trying to separate the financing timetable from current cash needs.
The headline number is stunning on its own, but the surrounding details show why the market is watching so closely. SpaceX’s Q3 revenue rose 161% from a year earlier to $14.58 billion, while AI revenue grew about 10 times to $7.56 billion, according to TipRanks. Capex nearly tripled to $19.6 billion in the same period. That combination of rapid top-line growth and even faster capital intensity is the core of the story: SpaceX is trying to scale an AI business that rents access to advanced chips, and it is doing so with a debt package large enough to move markets.
Musk is not just raising money. He is also using the moment to taunt competitors. In posts cited by Bloomberg and MarketWatch, he warned that “200+ neoclouds” face “guppy-level death and consolidation” and said SpaceX “is not buying any neoclouds. Pointless.” He also wrote, “As of the past few weeks, debt markets have almost entirely shut off access to money for any company that doesn’t have substantial revenue independent of AI data centers.” The message is as much strategic as it is financial: SpaceX wants to frame itself as one of the few players with enough non-AI revenue to survive a capital squeeze.
That argument is central to how investors may interpret the debt. SpaceX is leaning on Starlink and launch revenue to justify a financing structure that would likely be difficult for pure-play AI infrastructure names to replicate. CFO Bret Johnsen has tried to reinforce that point, telling investors, “We have a lot of conviction that we are going to be able to do this in a very measured fashion and continue to keep ourselves in that IG status along the way.” The company’s pitch is that it can expand aggressively while staying inside the lines of investment-grade discipline.
Still, the market is clearly not taking the message at face value. SpaceX credit-default swaps surged to a record 266 basis points at Friday’s close, a level typically associated with junk-rated risk. Separately, SpaceX 2056 bonds were trading near 85 cents on the dollar, with a yield about 2.27 percentage points over Treasuries, according to Financial Express citing FT-sourced data. Those are not the numbers of a company enjoying calm credit conditions. They point to a market that is pricing in size, complexity and uncertainty all at once.
The structure of the financing also matters. The deal is reported to involve Goldman Sachs as bankers, Apollo Global Management as lead investor, and Apollo-backed Atlas SP as entity manager, with Pimco considered as a potential investor. It is being issued as 144a private placements, a format usually reserved for large institutional buyers. That setup suggests this is not a simple public bond sale, but a carefully structured private-market transaction designed to move a massive amount of capital while preserving flexibility for the company and its backers.
Musk’s claim that the financing has already closed is not the end of the story, however. Bloomberg reported that pricing talks were still ongoing and that the deal had not priced as of late Sunday, Oct. 25, 2026. The Edge Malaysia also said final pricing and closure were expected in the coming days. For investors, that distinction matters. A claimed close and a priced, settled transaction are not the same thing, especially in a market where jumbo AI-linked financing can shift quickly if demand cools or terms need to change.
The scale of the package also raises a bigger question about SpaceX’s ambitions. The company is clearly not treating AI as a side project. It is building what looks like a major chip-intensive operation, with deliveries of GPUs scheduled for the first and second quarters of 2027. The timing suggests the debt is tied to a longer industrial ramp, not just a one-off purchase. In other words, SpaceX is preparing to absorb a large amount of hardware and use it to expand a business line that has already begun to matter materially in its financial mix.
The company’s next spending phase may be even heavier. Bret Johnsen guided Q4 2026 capex to $35 billion to $40 billion, according to TipRanks and Analytics India Mag. That guidance places the current debt deal inside a much larger investment cycle. If SpaceX is correct that it can keep its investment-grade standing while financing such rapid growth, it could become a template for other revenue-rich companies trying to straddle infrastructure, AI and credit markets. If it is wrong, the CDS spike may prove to be an early warning.
There is also a longer-dated strategic layer. TipRanks said the Terafab, a SpaceX-Tesla chip joint venture, is targeting its first test chip for May or June 2027, with volume production set for the third quarter of 2028. That timeline suggests SpaceX’s AI push is not limited to renting chips or buying compute. It is also building toward a more integrated hardware strategy. The debt package, then, is not just about one purchase. It is about financing a broader industrial stack that could take years to fully materialize.
For now, the market is left with a striking contrast: a founder publicly declaring victory on a $40 billion GPU financing, while credit investors mark the risk at record levels. SpaceX wants to be seen as one of the rare AI winners that can borrow big and still look safe. The next test is whether the deal is priced on the company’s terms, and whether the GPU bet turns into a business line that can justify the leverage it is taking on.