The past two weeks handed Nvidia’s skeptics their best material in months. CoreWeave had taken on roughly $51 billion of debt and lease obligations to finance its AI infrastructure, and it faced a rebellion from stock and bond investors as concerns mounted about AI overspending. The yield to maturity on CoreWeave’s 9.625% bond due in 2032 climbed to around 15%. Oracle bonds sat dangerously close to junk territory as markets cooled on artificial intelligence. The argument building in credit markets was straightforward: the AI order book was built on borrowed money that was getting harder to roll over.
Then SpaceX answered the question of who, exactly, is still buying.
Elon Musk said SpaceX had closed $40 billion in debt financing for graphics processing units, crediting the company’s investment-grade status and revenue from its Starlink satellite internet and launch businesses. The chips are expected to arrive in the first and second quarters of 2027, and Musk said the current quarter would not be significantly affected by the financing. The structure is mainly callable debt, meaning SpaceX can repay it early. Reports in early October said SpaceX was in talks with banks and investors to raise $40 billion, with some reporting a mix of roughly $10 billion in loans alongside $30 billion in investment-grade debt.
That is not a story about one company’s ambition. It is a story about Nvidia’s revenue calendar. A $40 billion GPU purchase committed to H1 2027 delivery lands precisely where doubts about forward demand were concentrating. Nvidia executives have said demand and purchase commitments for Blackwell and Rubin products could total $1 trillion through 2027. SpaceX’s close adds hard contract weight to a figure that critics had been treating as aspirational.
The credit angle is where the long-term investor should focus. Musk did not simply announce a deal; he drew a dividing line through the entire AI infrastructure market. He said that debt markets had “almost entirely shut off access to money for any company that doesn’t have substantial revenue independent of AI data centers.” He predicted “guppy-level death and consolidation” among more than 200 neocloud operators. SpaceX still got its deal done, he said, because “we’re investment grade and have Starlink and launch revenue.”
That distinction is not incidental to Nvidia’s demand story. It reshapes it. If that framing holds, the buyers who remain active in the GPU market are not the leveraged neoclouds. They are the hyperscalers with balance sheets and the diversified operators, like SpaceX, whose credit rests on revenues that predate the AI trade. Those are also the customers least likely to cancel.
The credit divide Musk described is already visible in broader debt markets. Global corporate bond spreads have blown out to their widest in six months, with AI infrastructure borrowers facing the sharpest scrutiny from credit committees.
Amazon’s AWS had already committed to receive roughly $5 billion of Nvidia GB300 GPU server racks upfront, with around $3 billion more per month expected for the foreseeable future. ByteDance, Alibaba, and Tencent had jointly ordered over $35 billion of Nvidia RTX PRO 5500 chips, though reporting on the deal has said Nvidia may not fully meet this demand due to capacity limits and export rules. SpaceX’s closed deal fits a pattern: the largest, most creditworthy entities are accelerating commitments while smaller, debt-dependent operators face a squeeze.
For Nvidia, the risk was never whether demand existed. The company’s worldwide production ceiling for RTX PRO 5500 has been reported at roughly 3 million cards for 2027, versus reported China orders of more than 6 million, which means booked demand does not automatically translate into shipped product or recognized revenue. Supply constraints, not order cancellations, are the more plausible constraint on earnings. A market that spent two weeks pricing in order-book collapse just received a $40 billion reminder that the buyers with the strongest balance sheets are still writing checks.
The investment-grade divide Musk described is a structural shift worth watching beyond a single weekend announcement. Lenders are no longer treating all GPU-backed debt as equivalent. Businesses with durable, non-AI revenue streams can still access capital to buy Nvidia hardware; businesses without them are facing terms that make further expansion very difficult. That bifurcation concentrates future demand at the top of the customer pyramid, which is exactly where Nvidia’s highest-margin relationships live.
