Musk says SpaceX has completed $40 billion in debt financing to purchase GPUs
Musk said on social media that SpaceX has completed $40 billion in debt financing, with the funds to be used to purchase GPUs, expected to be in place in the first and second quarters of 2027. He mentioned that SpaceX has an investment-grade rating and predicted that more than 200 new cloud service providers will face large-scale bankruptcy and consolidation. SpaceX's SI business revenue in the third quarter was $7.6 billion, mainly obtained by leasing data center capacity to Anthropic, Google, Reflection AI and others.
Why read this
AI score 75
Musk disclosed SpaceX's $40 billion GPU procurement financing arrangement and offered his judgment on oversupply in the computing power rental market.
Topics
Body · AI translation
In the early hours of October 26 Beijing time, Elon Musk posted on social media that SpaceX has completed $40 billion in debt financing, with the funds to be used to purchase GPUs. "This financing consists primarily of redeemable debt used for GPUs, and is expected to be in place in the first and second quarters of 2027," he added.
Musk noted that for companies without substantial revenue sources independent of AI data centers, the debt market has almost completely cut off their access to funding over the past few weeks. SpaceX, by contrast, has an investment-grade rating and generates revenue through Starlink and its rocket launch business. He also predicted that more than 200 new cloud service providers (Neoclouds) will face massive bankruptcies and consolidation, with only a few able to survive.
However, Musk's post did not specify what exactly "completed" refers to in the transaction, and it is currently unclear whether investors have reached agreement on the pricing of this debt.
There were earlier reports that SpaceX was in talks with banks and investors to raise $40 billion to buy Nvidia chips. The proposed financing would use a special purpose vehicle (SPV) to keep the debt off SpaceX's balance sheet, thereby preserving its investment-grade credit rating.
This financing also highlights the enormous capital demands of the AI boom. According to media reports, a Goldman Sachs report estimates that Amazon, Alphabet, Microsoft, Oracle and Meta will invest $800 billion in AI capacity building this year. The Goldman Sachs team believes spending will still grow substantially next year, but the pace of capital expenditure growth will begin to slow.
SpaceX also plans to step up its investment in AI. On the third-quarter earnings call, SpaceX Chief Financial Officer Bret Johnsen said the company expects fourth-quarter capital expenditure of $35 billion to $40 billion, with continued investment in computing power, Starship and satellite production lines.
It is worth noting that just a few days ago, SpaceX delivered its first profitable financial report since going public.
Driven by accelerated growth in its Super Intelligence (SI) and connectivity businesses, the company said it achieved third-quarter revenue of $14.6 billion, up 161% year over year; operating profit of $2.791 billion, net profit of $4.4 billion, and adjusted EBITDA of $8.809 billion. In the second quarter of this year, the company posted an operating loss of $143 million.
Among these, the AI business has become the biggest engine of SpaceX's revenue growth, surpassing the connectivity business including Starlink. In the third quarter of this year, SpaceX's SI business revenue was $7.6 billion, up 195% quarter over quarter and 830% year over year. This growth was mainly driven by the activation of new cloud service agreements, the acquisition of Cursor, and increased adoption of Grok by enterprises and consumers.
SI business revenue mainly comes from computing power leasing. The company earns AI revenue by leasing data center capacity to Anthropic, Google, Reflection AI and others.
Separately, Musk disclosed on the earnings call that SpaceX's computing power uses a dynamic allocation of 50% for internal use and 50% for external leasing, prioritizing short-term spot contracts, with a payback period of less than one year on computing power investment. "Major cloud providers and dozens of small and medium-sized enterprises keep approaching us hoping to lease computing power; demand far exceeds supply."