Key Takeaways
- SpaceX sold approximately $32 billion in debt to finance its purchase of advanced artificial intelligence chips, according to Bloomberg.
- The transaction was completed in three parts, with the largest portion consisting of approximately $25.2 billion of Class A3 notes carrying a yield of 8.259%.
- The financing forms part of a broader agreement worth approximately $40 billion, arranged with Apollo Global Management and Goldman Sachs.
- The structure uses a special-purpose vehicle to purchase Nvidia chips and lease them to SpaceX, helping keep the borrowing separate from the company’s direct balance sheet.
- The chips are expected to support SpaceX’s terrestrial AI computing infrastructure, including its plans to expand computing capacity for xAI’s Grok products and other customers.
- Investors are demanding relatively high returns despite SpaceX’s investment-grade credit profile, reflecting concerns about the risks and complexity of AI infrastructure financing.
- The transaction could become a model for other technology companies seeking to finance expensive GPU purchases without directly carrying all the associated borrowing on their balance sheets.
SpaceX has raised approximately $32 billion through a major debt sale to finance its artificial intelligence infrastructure expansion, marking one of the largest financing transactions linked to AI chips. The Elon Musk-led company sold the debt on November 19, 2026, through a complex financing structure involving Apollo Global Management and Goldman Sachs. The transaction is part of a broader financing package of more than $40 billion designed to fund the acquisition of advanced Nvidia processors and related server equipment. The debt was sold in three portions, with the largest tranche carrying a yield of 8.259%, according to Bloomberg.
The financing highlights the enormous capital requirements of the global AI industry. Technology companies are spending billions of dollars to secure advanced processors, build data centres and develop the infrastructure required to train and operate increasingly powerful AI systems. SpaceX is using a special-purpose vehicle to purchase the chips and lease them to the company for approximately five years. The arrangement is designed to keep the borrowing off SpaceX’s direct balance sheet and help preserve its investment-grade credit profile. However, the yields investors are demanding show that even a highly valued technology company must pay a significant price to finance the AI computing boom.
SpaceX Completes a $32 Billion AI Chip Debt Sale
SpaceX sold approximately $32 billion of debt on Thursday, November 19, as part of its broader AI chip financing programme.
According to Bloomberg, the transaction was divided into three portions and arranged by Goldman Sachs and Apollo Global Management.
The largest portion consisted of approximately $25.2 billion of Class A3 notes, which were priced at a yield of 8.259%.
That yield is significant because it is considerably higher than the average borrowing cost for investment-grade corporate bonds at the time.
Bloomberg reported that the average yield on US high-grade corporate notes was approximately 5.95% on the day of the transaction.
The difference indicates that investors are seeking a substantial premium for financing the equipment-backed structure, even though SpaceX has an investment-grade credit profile.
The deal also follows concerns about an earlier SpaceX unsecured bond offering, whose securities reportedly declined in the secondary market after trading began.
That experience appears to have made investors more cautious about the company’s borrowing arrangements.
Key details of the debt transaction
| Transaction detail | Reported information |
|---|---|
| Company | SpaceX |
| Debt sold | Approximately $32 billion |
| Broader financing programme | More than $40 billion |
| Largest debt tranche | Approximately $25.2 billion |
| Yield on largest tranche | 8.259% |
| Number of portions sold | Three |
| Main transaction arrangers | Goldman Sachs and Apollo Global Management |
| Financing vehicle | Saturn Funding Trust |
| Equipment | Nvidia AI processors and related hardware |
| Lease period | Approximately five years |
The $32 billion debt sale is not the entire financing programme. It forms part of a broader arrangement that also includes a senior secured loan and other financing commitments.
The overall package is intended to support approximately $47.3 billion of GPU equipment purchases.
Why SpaceX Is Borrowing $40 Billion for Nvidia Chips
The financing is connected to SpaceX’s ambition to build substantial AI computing infrastructure.
Advanced graphics processing units, or GPUs, are among the most expensive and strategically important components of modern AI infrastructure.
They are used to train large AI models, process complex datasets and operate AI applications at scale.
Companies developing advanced AI systems need access to thousands of processors working together in specialised data centres.
SpaceX is seeking to secure a large supply of Nvidia’s next-generation Vera Rubin computing hardware.
The equipment is intended to support its terrestrial computing infrastructure, including AI workloads associated with xAI’s Grok products and cloud computing services.
SpaceX announced in October that it had reached a binding commitment to enter into an approximately $40 billion master finance lease with an Apollo-managed entity.
The agreement covers GPUs and related equipment, with deliveries expected during the first half of 2027.
The company has argued that the financing will help it deploy computing infrastructure quickly while diversifying the sources of capital available for its expansion.
Why AI companies need so much capital
Building an AI data centre involves much more than purchasing processors.
Companies must also invest in:
- High-performance servers and networking equipment.
- Electricity generation and power infrastructure.
- Advanced cooling systems.
- Data-centre buildings and land.
- High-speed connections between computing facilities.
- Software and engineering teams.
- Maintenance and replacement of expensive hardware.
A large AI computing project can therefore require tens of billions of dollars before it reaches full operating capacity.
SpaceX’s financing demonstrates how companies are increasingly turning to specialised debt structures to fund these investments.
How the Apollo and Goldman Sachs Financing Structure Works
One of the most important aspects of the SpaceX transaction is the way the debt has been structured.
Instead of SpaceX directly purchasing all the chips using its own cash or borrowing the entire amount on its balance sheet, the financing uses a special-purpose vehicle.
The vehicle is known as Saturn Funding Trust.
It is managed through an Apollo-backed entity and is designed to purchase the Nvidia processors and related equipment.
SpaceX then leases the equipment from the vehicle for approximately five years.
The structure separates the financing of the physical equipment from SpaceX’s direct corporate borrowing.
How the financing works
- Investors provide capital: Institutional investors purchase debt securities issued through the financing structure.
- The financing vehicle purchases equipment: Saturn Funding Trust uses the funding to acquire Nvidia GPUs and related hardware.
- SpaceX leases the equipment: The company obtains access to the processors through a lease arrangement.
- SpaceX makes lease payments: The payments support the financing vehicle’s obligations to investors.
- Investors receive returns: Debt holders receive interest and principal according to the securities’ terms.
<AsyncImage query=”Nvidia AI GPU server racks hyperscale data center close up” aspectRatio=”16:9″ width=”100%”/>
The structure is intended to help SpaceX obtain substantial computing capacity without taking on the entire financing package as conventional corporate debt.
However, describing the transaction as off-balance-sheet financing does not mean SpaceX has no financial obligation.
The company has committed to making lease payments under the arrangement. Bloomberg reported that SpaceX remains responsible for payments even if the equipment fails or an extraordinary event, such as a natural disaster, occurs.
The actual accounting treatment depends on the applicable accounting standards and contractual terms. The structure should therefore not be interpreted as eliminating the company’s economic exposure.
Why Investors Are Demanding Yields Above 8%
The 8.259% yield on the largest tranche is one of the most important features of the transaction.
Investors generally demand higher yields when they believe a security carries additional credit, liquidity, structural or valuation risk.
SpaceX’s debt sale comes at a time when investors are becoming more selective about financing AI infrastructure.
The industry has attracted enormous investment, but concerns are increasing about how quickly companies can generate returns from expensive computing assets.
1. AI infrastructure requires massive investment
The cost of acquiring GPUs and building data centres has increased substantially.
Companies are committing billions of dollars to infrastructure before knowing exactly how quickly demand will translate into sustainable cash flow.
Investors must therefore assess whether future AI revenue will justify the enormous upfront spending.
2. GPU values can change quickly
AI processors are expensive assets, but their commercial value depends on performance, availability and demand.
New generations of processors can make older equipment less attractive for certain workloads.
Investors financing GPUs must consider how much the equipment could be worth if the borrower defaults or the lease ends.
3. The financing structure is complicated
The transaction involves a special-purpose vehicle, equipment ownership, lease obligations and multiple categories of debt.
These arrangements can be more difficult for conventional corporate bond investors to evaluate than straightforward unsecured bonds.
4. Investors are becoming cautious about AI debt
The rapid expansion of AI infrastructure has created concerns about excessive borrowing.
Several companies are financing large data-centre projects through debt, equipment leases and other structured arrangements.
If AI demand grows more slowly than expected, companies could face difficulties generating enough cash flow to support their financing obligations.
The yield premium therefore reflects more than SpaceX’s own credit profile. It also reflects the risks associated with the wider AI infrastructure financing market.
SpaceX’s Debt Costs More Than Its Existing Corporate Bonds
The new financing is relatively expensive compared with SpaceX’s earlier borrowing.
Bloomberg reported that the average US investment-grade bond yield was approximately 5.95% on November 19.
The 8.259% yield on the largest SpaceX-backed tranche was therefore approximately 2.31 percentage points higher.
The securities also offered a premium compared with SpaceX’s existing debt from its June bond offering.
This matters because the company is not simply borrowing money at the same rates available to established investment-grade businesses.
Investors are demanding additional compensation for the specialised financing structure and the risks connected to AI hardware.
The transaction also follows SpaceX’s earlier unsecured bond sale, which reportedly performed poorly in the secondary market after issuance.
The new equipment-backed structure may provide investors with additional protections, but its complexity means that buyers still need to assess the risks carefully.
SpaceX Wants to Protect Its Investment-Grade Credit Rating
Maintaining an investment-grade credit profile is important for a company planning to borrow billions of dollars.
Investment-grade borrowers generally have access to a wider pool of institutional investors and can often secure financing on more favourable terms than companies with speculative-grade ratings.
SpaceX has been able to present itself as more than an AI infrastructure business.
Its operations include rocket launches and Starlink satellite internet services, which provide revenue sources beyond AI data centres.
Chief Financial Officer Bret Johnsen has previously highlighted the company’s investment-grade balance sheet.
The financing structure is designed to help preserve that position by keeping the equipment-related borrowing separate from conventional corporate debt.
That does not mean the arrangement is risk-free.
SpaceX still has substantial lease-payment commitments, and investors will examine whether its existing businesses can support the company’s expanding capital requirements.
The company’s ability to generate cash from launch services, Starlink and AI computing will therefore remain important to its long-term financing strategy.
The Financing Could Become a Model for the AI Industry
SpaceX’s transaction is significant because it demonstrates how specialised financing can help technology companies acquire expensive AI hardware.
Traditional corporate borrowing is not the only way to finance computing infrastructure.
Companies can also use equipment leases, asset-backed securities, special-purpose vehicles and other structured financing arrangements.
These structures allow investors to finance specific assets rather than relying entirely on the creditworthiness of the company purchasing the equipment.
The model could become increasingly important as AI infrastructure spending expands.
According to estimates cited by industry participants, companies may need approximately $1.5 trillion to finance the chips required for the broader AI buildout.
If that demand materialises, banks, private-credit firms, asset managers and institutional investors could see more opportunities to finance GPUs and data-centre equipment.
However, the market is still relatively new.
Investors must understand how quickly hardware loses value, how lease payments are protected and whether equipment can be recovered and resold if a borrower defaults.
SpaceX’s transaction could help establish a benchmark for future GPU financing deals.
Concerns About SpaceX’s Potential Orbital Data Centres
The financing has also attracted attention because Elon Musk has discussed the possibility of placing AI data centres in space.
This creates a potential issue for lenders financing expensive Nvidia processors.
If chips used as collateral were sent into orbit, recovering and selling them after a default could become extremely difficult.
The Financial Times reported on November 20 that lenders involved in the broader SpaceX financing package sought to prevent the company from launching the financed chips into orbit.
The concern is straightforward: creditors need to be able to recover assets that secure their financing.
Equipment located in a conventional terrestrial data centre can potentially be repossessed, transferred or sold.
Hardware operating in orbit presents a much more complicated recovery challenge.
The Financial Times said it was unclear whether the proposed restriction made it into the final loan documents.
SpaceX has said the chips financed through the arrangement would be used in US data centres that it majority-owns, controls and operates.
This distinction is important. Musk’s longer-term interest in orbital computing does not establish that the equipment covered by this particular financing will be deployed in space.
What the Deal Means for Nvidia
Nvidia is another important beneficiary of the financing.
The company supplies the processors that power many of the world’s most advanced AI systems.
As technology companies expand their computing infrastructure, demand for Nvidia’s latest processors has become a major driver of data-centre investment.
SpaceX’s financing is designed to fund a large purchase of Nvidia hardware, including its next-generation Vera Rubin systems.
For Nvidia, large financing commitments can help customers secure expensive equipment and accelerate deployment.
However, the transaction also highlights a potential challenge for the semiconductor industry.
Demand for advanced processors increasingly depends on whether customers can secure sufficient capital to purchase them.
If financing becomes more expensive or investors become less willing to support AI infrastructure projects, some companies could delay equipment purchases.
This means the cost and availability of capital may become just as important as chip supply in determining the pace of AI infrastructure expansion.
The Bigger Picture
SpaceX’s $32 billion debt sale demonstrates how the AI boom is creating a new market for specialised equipment financing. Instead of relying exclusively on conventional corporate borrowing, companies are increasingly exploring structures that connect investors directly to the equipment required for AI computing. The use of special-purpose vehicles and leasing arrangements could allow technology companies to secure expensive processors while managing the impact of borrowing on their reported balance sheets.
However, the transaction also highlights the risks behind the AI investment cycle. An 8.259% yield on the largest tranche shows that investors are not treating every AI infrastructure investment as a low-risk opportunity. GPU depreciation, technological obsolescence, complicated financing structures and uncertain future demand all influence the returns investors require. As more companies seek billions of dollars to build AI infrastructure, the ability to finance these investments at sustainable costs could become a major competitive advantage.
Looking Ahead
The next important development will be the closing of the broader financing package and the delivery of Nvidia’s next-generation processors during the first half of 2027. SpaceX will need to deploy the equipment, generate sufficient demand for its computing capacity and meet the lease obligations associated with the financing. Investors will also watch how the company balances AI infrastructure spending with its existing launch and Starlink businesses. The transaction’s final performance will depend not only on the availability of advanced chips but also on whether the computing capacity generates enough revenue to justify its cost.
For the wider technology industry, SpaceX’s financing could mark an important step in the development of a dedicated market for AI hardware-backed debt. Banks, private-credit firms and institutional investors may increasingly finance GPUs through equipment leases and asset-backed structures. But the high yields and complex terms show that this market is not without risk. As AI infrastructure investment continues to grow, the cost of financing, the resale value of processors and the ability of companies to turn computing capacity into reliable revenue will determine whether this model becomes a lasting part of the global AI economy.
Get the day’s top stories in your inbox
One concise email. No spam, unsubscribe anytime.



