SpaceX’s Chip Financing Talks Put AI Debt Appetite to the Test

SpaceX is in early-stage talks to raise or borrow $40 billion to buy Nvidia chips, Ed Ludlow reported. The proposed financing has two parts: $10 billion in loans and $30 billion in investment-grade credit. The distinction matters to the structure of the plan, but Ludlow did not describe the loans’ terms or say how the credit would be arranged. He did say the scale of the borrowing had drawn attention to warning signs, including a credit-default-swap chart that appeared after the news. He did not explain what the chart showed.

The market’s willingness to support the financing sits alongside those concerns, not in place of them. Ludlow pointed to a debate about the useful economic life of a GPU and the best way to finance one so it can be put to work in a data center quickly. The financing question is therefore not only how much capital a company can raise, but how the cost and timing of that capital fit the period in which the chips are expected to be useful. Ludlow did not give a view on how long that period is or how the proposed financing would be repaid.

The deal also sharpens a question about what comes after a major IPO. Months before a possible SpaceX IPO, Ludlow said, investors and bankers had discussed the prospect that the company would go to the capital markets afterward. That was an expectation being discussed, not an IPO that had already happened. The current borrowing talks put the follow-on financing question in concrete terms: might investors be asked to support a large public offering and then a substantial raise in debt markets?

The host asked whether SpaceX might offer a blueprint for Anthropic and OpenAI, which had yet to come to market. Ludlow said investors considering those eventual IPOs may need to be comfortable with the possibility that the companies would turn to debt markets soon afterward. But he cautioned against treating SpaceX as the origin of the pattern. Cloud-computing companies that build and operate their own data centers—the hyperscalers—had already been borrowing at unprecedented levels for about 18 months, he said.

The borrowing is only part of what investors have been accepting. Ludlow said the market has also been relatively relaxed about the possibility that these companies could move into negative free cash flow as a consequence of their spending and borrowing. That context makes the proposed SpaceX financing less a wholly new template than another instance of a broader willingness to fund data-center investment with substantial borrowing, even when spending pressures cash flow. Whether investors would accept the same pattern from companies coming to market for the first time remains the question the host raised.

Apple is looking to work with LG on smart-home products including doorbells, locks and thermostats. Ludlow described the arrangement as a joint venture, which he said is relatively unusual for Apple. He framed it as part of a deeper push into the home, with Apple drawing on LG’s existing expertise in product lines such as smart locks and doorbells. The host asked whether the products would carry Apple branding, but Ludlow did not answer that point directly.

The on-screen material included technical line drawings attributed to the FCC. They show the backplate and battery compartments of a product labeled “LG Smart Deadbolt Lock,” and a circular thermostat design. Those images establish what the displayed drawings depict; they do not, by themselves, establish the final product lineup or how the products would be branded. Ludlow also referred viewers to diagrams in a story by Mark Gurman.

The host’s concern was whether Apple might be late to a market where many households have already chosen their devices and may have little reason to switch. Ludlow answered from his own experience rather than from a reporting standpoint. His home combines a SimpliSafe security system and Amazon Echo devices with an iPhone and MacBook. The mix illustrates the practical challenge he described: a customer can use Apple products without having chosen Apple for the rest of the home.

Apple’s bet, as Ludlow characterized it, is that iPhone or Mac users may see value in bringing more of their hardware together. LG’s role would draw on expertise in existing product categories, while the appeal to customers would depend on whether the resulting products fit well enough into the Apple hardware they already use. Those are related but distinct parts of the proposition: LG contributes experience in the categories, and Apple is counting on the value of a more unified set of devices.

Ludlow called the market “tough and saturated.” His example points to why: a household may already have a security system and smart-home devices from one company while relying on Apple for phones and computers. Apple would need to persuade customers that adding or changing home devices is worth doing, rather than assuming that owning an iPhone or Mac will be enough to overcome established habits.