JupiteX Get the app
Markets & Finance21 Nov 2026 · about 7 min

Apollo, Goldman lead $44 billion SpaceX GPU debt financing

The brief

The financing is intended to fund GPUs and related equipment for SpaceX. GPUs are advanced processors, and the equipment around them helps create computing infrastructure. The deal reflects the large sums technology companies are committing as they race to expand capacity for AI development. One specific target is the next generation of Nvidia Vera Rubin chips. The transaction is expected to finance these chips for SpaceX’s terrestrial compute buildout. The financing consists of senior and subordinated notes, which Apollo and Goldman Sachs said they priced at about $44 billion. The article does not list all the related equipment being purchased. The deal was oversubscribed and increased in size after strong investor demand. Delivery is expected in the first half of 2027, when lease schedules are also expected to start. The financing supports SpaceX’s planned computing buildout, but the article does not specify the system’s final size or exactly what work it will perform.

01

What equipment is the $44 billion financing intended to buy for SpaceX?

The financing is intended to fund GPUs and related equipment for SpaceX. GPUs are advanced processors, and the equipment around them helps create computing infrastructure. The deal reflects the large sums technology companies are committing as they race to expand capacity for AI development.

One specific target is the next generation of Nvidia Vera Rubin chips. The transaction is expected to finance these chips for SpaceX’s terrestrial compute buildout. The financing consists of senior and subordinated notes, which Apollo and Goldman Sachs said they priced at about $44 billion. The article does not list all the related equipment being purchased.

The deal was oversubscribed and increased in size after strong investor demand. Delivery is expected in the first half of 2027, when lease schedules are also expected to start. The financing supports SpaceX’s planned computing buildout, but the article does not specify the system’s final size or exactly what work it will perform.

02

How large is this deal, and what makes it the largest GPU financing to date?

The deal totals about $44 billion in senior and subordinated notes. Apollo and Goldman Sachs said they priced the notes to fund GPUs and related equipment for SpaceX. The financing matters because it shows the scale of investment required as technology companies race to secure advanced processors and build computing infrastructure for AI development.

The firms described it as the largest GPU financing to date. The transaction was oversubscribed, meaning investor demand exceeded the amount initially offered, and it was upsized in response. The article names BlackRock, Goldman Sachs Asset Management and accounts advised by KKR as anchor commitments.

The article does not give a ranked comparison with other GPU financings, so the basis for the “largest” description beyond the deal’s scale is not detailed. It does say that strong investor demand helped expand the transaction. The funds are expected to support SpaceX’s terrestrial compute buildout, including next-generation Nvidia Vera Rubin chips.

03

What are senior and subordinated notes, and how does their repayment priority differ?

Senior and subordinated notes are types of debt. The key difference is their repayment order if a borrower cannot meet its obligations: senior noteholders generally have priority over subordinated noteholders. This describes a common feature of debt, not specific terms disclosed for this SpaceX financing.

For example, if available funds were insufficient to repay all creditors, senior debt would typically be paid before subordinated debt. Subordinated noteholders would receive payment only after higher-priority claims were addressed. The precise order can depend on the contracts and other obligations, details the article does not provide.

Apollo and Goldman Sachs said they priced about $44 billion of senior and subordinated notes to fund GPUs and related equipment for SpaceX. The article does not state how much of the financing is in each category, the interest rates, or the detailed repayment terms. So the general priority difference can be explained, but the exact protections for investors in this deal cannot be determined from the reported details.

04

How will the lease arrangement work, and when are payments expected to start?

The arrangement separates buying the equipment from using it. SpaceX said an affiliate of Apollo Global Management would buy AI chips and server hardware, then lease that equipment to SpaceX. The article gives total principal payments of about $40 billion for that arrangement, while the separate financing announced by Apollo and Goldman Sachs is about $44 billion.

In a lease, the company using equipment makes payments for the right to use it rather than buying it directly. Here, the Apollo affiliate is the stated buyer and lessor, and SpaceX is the intended lessee. The article says the notes will fund GPUs and related equipment, but it does not detail lease rates or payment dates beyond when schedules are expected to begin.

Lease schedules are expected to start when the GPUs and related equipment are delivered in the first half of 2027. That means the start is tied to delivery, not a date specified more precisely in the report. The article does not say when delivery will finish or how long the leases will run.

05

What will the financed GPUs and related equipment enable SpaceX to do?

The financed GPUs and related equipment are meant to expand SpaceX’s computing infrastructure. That matters because the article describes technology companies racing to secure advanced processors and build out computing capacity needed to support AI development. It does not say exactly which products or services SpaceX plans to run on the equipment.

A concrete example is the expected financing of the next generation of Nvidia Vera Rubin chips for SpaceX’s terrestrial compute buildout. The article also describes an Apollo affiliate buying AI chips and server hardware and leasing them to SpaceX. The mechanism is therefore a combination of equipment financing and leasing, rather than a disclosed plan for a specific AI application.

Once delivered, the equipment is expected to add computing capacity on the ground. Lease schedules are expected to start with delivery in the first half of 2027. The article does not give a capacity figure, launch date for any AI service, or details about how SpaceX will use the computers, so those outcomes cannot be stated from the reported facts.

06

How can leasing costly equipment help a company build computing capacity without buying all of it upfront?

Leasing is one way for a company to obtain use of costly equipment without buying all of it directly at the outset. Instead, another party buys the equipment and lets the company use it in return for payments. This can help a company add computing capacity while structuring the cost as lease obligations; leasing does not mean the equipment is free or that costs disappear.

The article describes this kind of arrangement for SpaceX. An affiliate of Apollo Global Management is expected to buy AI chips and server hardware and lease them to SpaceX. SpaceX said last month that the deal had total principal payments of about $40 billion. Separately, Apollo and Goldman Sachs priced about $44 billion in notes to fund GPUs and related equipment.

Lease schedules are expected to start when the GPUs and related equipment are delivered in the first half of 2027. The article does not spell out the lease’s duration, payment calendar, or how much SpaceX would have paid upfront under a direct purchase. It does show how a leasing structure is being used to support a large computing buildout.

07

What is a GPU, and why is it useful for the calculations involved in developing and running AI?

A GPU, or graphics processing unit, is a type of processor that can perform many calculations in parallel. Although GPUs were developed for graphics work, this parallel processing also makes them useful for the large mathematical workloads involved in AI. This definition comes from general computing knowledge; the article itself does not explain what a GPU is.

AI development and use can involve repeated calculations on large amounts of data. GPUs can perform many such operations at the same time, which is why companies seek advanced processors and computing infrastructure for AI. The article says technology companies are racing to secure these processors, but it does not describe the specific calculations SpaceX plans to run.

The financing is expected to fund GPUs and related equipment, including the next generation of Nvidia Vera Rubin chips for SpaceX’s terrestrial compute buildout. The article places this in the broader context of investment in AI infrastructure. It does not provide performance figures for the chips or say which AI models or applications will use them.

This brief was written by AI from the original reporting and checked by other models. Names, figures and quotes come from the source; read it for full context.

Read more in the JupiteX app

Pulse is free. New stories every 4 hours, each one broken into the questions that explain it.

Or read more news on the web