|
Getting your Trinity Audio player ready...
|
SpaceX is reported to be lining up one of the largest debt packages ever raised to pay for artificial intelligence hardware. According to the Financial Times, which broke the story on Tuesday, and Bloomberg, which followed with its own account, Elon Musk’s company is talking to banks and investors about raising $40 billion to buy chips from Nvidia. Both outlets cite people familiar with the matter, and neither SpaceX nor Nvidia is quoted in the reports reviewed, so the plan remains unconfirmed.
The structure reported so far is fairly specific. Reuters, relaying the Financial Times, says the company is seeking about $10 billion in bank loans and $30 billion in investment-grade debt, with the asset manager Apollo Global Management leading the effort. Pacific Investment Management Co., better known as Pimco, is among the lenders said to be in discussions, according to both the Financial Times and Bloomberg. Bloomberg described the talks as being at an early stage and cautioned that they could end without an agreement, which is a useful reminder that a headline figure is not a signed deal.
If it goes ahead, the size is the story. Bloomberg characterized the package as among the biggest debt financings tied to the AI buildout, a field where borrowing has already reached levels that would have seemed implausible a few years ago. Chips are the most expensive single input in that race, because training and running advanced models requires vast clusters of graphics processors, along with the power and cooling to run them. The reported order would feed SpaceX’s data centers, and Musk said in August that the company would build them using only Nvidia hardware. Last month he added that the Colossus 2 site could more than double its Nvidia chip count by December.
The AI business sitting inside SpaceX is a recent development. According to publicly compiled records, SpaceX acquired Musk’s artificial intelligence company xAI in an all-stock deal in February, valuing xAI at $250 billion, and the unit was rebranded SpaceXAI in July. That means a launch and satellite company is now also one of the most aggressive buyers of AI computing in the world, with a rocket business, a satellite internet business and an AI lab under one roof. Borrowing to finance the chip purchases fits a company that wants to expand quickly without issuing large amounts of new stock, though the reports do not say why debt was chosen over equity.
There is also a commercial reason to buy so many chips: SpaceX does not only use them itself. The Next Web reported that the company earns money from the hardware by renting it to other firms, and that Google has agreed to pay $920 million a month for access to about 110,000 Nvidia GPUs. That figure comes from press reports, and the details of the arrangement have not been published by the companies involved in the coverage reviewed. If accurate, it would imply that each chip can generate substantial recurring revenue, which is the logic lenders will test when they decide whether to extend credit against it.
Lenders will also weigh the borrower. One report said SpaceX has obtained a BBB credit rating, which would open the door to institutional investors with strict mandates, such as pension and insurance funds. That claim comes from a single outlet, and the rating agencies’ own announcements were not part of the coverage reviewed, so it should be treated cautiously. Investment-grade status matters because it lowers borrowing costs and widens the pool of buyers, and a $30 billion bond placement would need deep demand from exactly that kind of investor.
The market reaction was muted but telling. SpaceX shares, which trade on Nasdaq under the ticker SPCX, fell about 1 percent in after-hours trading on Tuesday, while Nvidia rose roughly 0.5 percent, according to The Next Web. A separate report said SpaceX slipped about 2 percent in premarket trading on Wednesday. Small moves like these suggest investors see the financing as plausible and largely expected, though a stock decline on news of heavy borrowing is a common reaction because it signals higher interest costs and added risk.
The story also highlights how tangled the financing of AI has become. Nvidia has been helping customers pay for its products, and Benzinga noted that Apollo is already involved in financing platforms Nvidia has helped assemble with Apollo, BlackRock and Goldman Sachs, with the aim of mobilizing more than $500 billion for AI infrastructure. In that picture, lenders provide money, customers buy chips, and the chipmaker’s growth supports the confidence that makes more lending possible. Supporters say that is simply how capital-intensive industries grow, much as railways and telecom networks were financed in earlier eras. Critics worry about circularity, where the same flow of money props up demand and valuations at every step.
That concern echoes arguments made elsewhere this month. Investor Michael Burry, famed for betting against the housing market before 2008, has argued that debt-fueled spending on chips and data centers is vulnerable to higher interest rates and could unravel if spending slows, according to coverage of his recent remarks. He holds bearish positions, so his view should be read with that in mind, but the question he raises is the one every lender in a $40 billion deal has to answer. Will the revenue from AI services arrive in time to cover the interest on the chips? Bears say the bill comes first and the revenue later, while bulls point to demand that, in Bloomberg’s words as reported by other outlets, is insatiable.
For Nvidia, a customer willing to borrow tens of billions to buy more hardware is another sign that demand has not cooled. The company has posted record-level share prices this week, according to market coverage, and any large new order supports its growth story. It also deepens the dependence of the AI sector on a single supplier. A company that commits to buying only Nvidia hardware gains consistency and software compatibility but gives up some bargaining power, and it ties its fortunes to one chipmaker’s production schedule and pricing.
For everyone else, the practical effects are less direct but still real. The scale of AI infrastructure spending influences electricity demand, chip supply for other industries, and the cost of capital for companies competing for loans. It also shapes which firms can afford to train frontier models, and whether smaller players are priced out. Readers in emerging markets, including Nigeria, may feel the effects through the cost and availability of cloud computing and AI services over time, though that link is indirect and depends on how these investments pay off.
What to watch next is straightforward. Look for a confirmation or denial from SpaceX, the formal launch of a loan or bond sale, the final split between bank debt and bonds, the names of the lenders, and the pricing investors demand, which will show how the market views the risk. The companies’ own channels, including the SpaceX and Nvidia websites, will carry official updates if a deal is announced.
Readers who follow global technology and business developments can find more coverage at BusinessTech Nigeria. For now, the safe reading is that SpaceX is exploring an enormous debt raise to keep its AI buildout running, and that the world will soon learn whether lenders are ready to fund it.
