SpaceX Secures $40 Billion AI Hardware Financing Arrangement With Apollo

SpaceX has secured a binding commitment for an approximately $40 billion financing arrangement to acquire artificial intelligence chips and related data center equipment through a leasing structure managed by Apollo Global Management.

Announced on October 26, 2026, the agreement is designed to help fund SpaceX’s expanding computing infrastructure, with deliveries of NVIDIA’s next-generation Vera Rubin chips expected to begin in the first half of 2027.

The transaction is structured as a master finance lease rather than a conventional arrangement in which SpaceX directly borrows the full amount to purchase the hardware.

An Apollo-managed entity will acquire the graphics processing units (GPUs) and related equipment, then lease them to SpaceX.

The company expects the agreement to close before the end of 2026, subject to specified conditions.

SpaceX described the transaction as the largest committed GPU financing arrangement.

The approximately $40 billion figure represents expected aggregate principal payments under the leases; interest and certain additional costs are separate.

The financing therefore represents a substantial long-term commitment to computing infrastructure, not a $40 billion cash payment made immediately to SpaceX.

How the $40 Billion Financing Arrangement Works

Under the agreement, the Apollo-managed entity will finance purchases of GPUs and associated equipment, with hardware delivered in installments.

Individual lease schedules are expected to begin as the equipment arrives during the first half of 2027.

According to reporting on SpaceX’s regulatory filing, each lease is expected to run for slightly more than five years from the delivery date of the relevant equipment.

The leases will carry interest, while the final pricing and specific costs depend on the terms set before the leases begin.

SpaceX has not publicly disclosed every commercial detail in its announcement.

This structure gives SpaceX access to expensive AI hardware while distributing payments over the period in which the equipment is used.

It also creates a financing obligation tied to assets that can lose value as newer chips become available.

The arrangement’s financial impact will depend on the final lease terms, equipment utilization, operating costs, and the revenue generated by the computing capacity.

The financing commitment is not the same as a completed transaction.

SpaceX said the agreement remains subject to closing conditions and is expected to close before year-end 2026.

The scheduled start of lease payments is linked to delivery of the equipment in 2027.

Why SpaceX Is Investing So Heavily in AI Computing

The financing reflects SpaceX’s effort to expand beyond its established space and satellite operations into large-scale AI computing.

The company plans to use the additional infrastructure to support Grok, the AI service associated with xAI, as well as enterprise customers and other cloud computing services.

Graphics processing units are specialized processors that can handle the parallel calculations required to train and run many AI models.

Large AI systems need substantial numbers of these chips, together with servers, networking equipment, electricity, cooling systems, and suitable data center facilities.

Securing financing for the processors is one part of building that capacity; the equipment must still be delivered, installed, powered, and operated.

SpaceX Chief Financial Officer Bret Johnsen said the lease would diversify the company’s funding sources and help meet part of the capital required for its expansion.

In the company’s announcement, Johnsen also said SpaceX was seeing payback periods of less than one year on its AI infrastructure investments.

That is management’s characterization of its investment economics, rather than an independently verified return applicable to every facility or project. 

The company’s computing ambitions are also connected to its broader business strategy.

Revenue from satellite internet and launch services gives SpaceX businesses outside AI computing, while the company seeks to build infrastructure that can serve both its own AI products and external customers.

The new lease adds a financing route for that expansion, but it does not eliminate the operational and commercial risks involved.

Apollo, Goldman Sachs, BlackRock, and KKR Participate in the Financing

Apollo Global Management is central to the arrangement because an entity it manages will acquire the equipment and lease it to SpaceX.

Goldman Sachs served as the sole structuring agent, while funds or accounts managed by BlackRock and Goldman Sachs Asset Management, along with capital accounts advised by KKR, made anchor commitments, according to the companies’ announcement. 

The involvement of multiple financial institutions illustrates how capital markets are helping fund the infrastructure behind the AI industry.

Instead of requiring SpaceX to fund the entire equipment purchase directly at the outset, the lease structure places the acquisition with the financing entity and spreads SpaceX’s payments over the relevant lease periods.

However, the $40 billion headline should not be interpreted as $40 billion in financing proceeds paid directly into SpaceX’s operating accounts.

The announced amount refers to expected aggregate principal payments over the lease schedules.

Interest and certain additional costs are separate, and the final interest rate was still to be determined before the master lease was entered into.

This distinction matters when evaluating the transaction.

The arrangement provides access to equipment and a structured payment schedule, but it also creates a substantial future payment obligation.

The final economics will depend on the executed agreements, financing costs, delivery schedule, and the performance of the infrastructure.

NVIDIA’s Vera Rubin Chips Are Central to the Buildout

The lease is intended to provide SpaceX with NVIDIA’s next-generation Vera Rubin chips and related equipment for its terrestrial computing infrastructure.

NVIDIA develops GPUs and other hardware used in AI training and inference, the process through which trained models generate responses or perform tasks.

The hardware is expected to arrive in installments during the first half of 2027.

Delivering equipment in tranches allows separate lease schedules to commence as individual portions of the infrastructure arrive, rather than requiring the entire deployment to happen at once.

The announcement does not establish how many individual GPUs SpaceX will receive, the final performance of the completed systems, or how much computing capacity will become operational on each delivery date.

Those details should not be inferred from the financing amount alone.

Hardware availability is only one part of the deployment challenge.

Data centers also require adequate electrical power, cooling, networking, physical space, and reliable operations.

The pace at which SpaceX can convert financed equipment into usable computing capacity will depend on these requirements as well as the delivery and installation schedule.

What the Deal Means for the AI Infrastructure Market

SpaceX’s agreement highlights the scale of capital required to compete in advanced AI computing.

Companies building large AI systems need to secure increasingly sophisticated hardware while also financing the facilities and energy infrastructure needed to operate it.

Equipment leasing can help align payments with the useful period of the assets and reduce the need for a company to fund all equipment purchases immediately.

But it is not free capital.

The lessee must meet its contractual obligations even if demand weakens, hardware becomes less competitive, or the revenue generated by the infrastructure falls short of expectations.

That risk is particularly relevant in AI, where new generations of processors can change the economics of existing data centers.

A business that commits to multi-year payments needs enough demand and operating efficiency to support the cost of its equipment over the lease term.

SpaceX’s arrangement also provides a useful distinction between companies with multiple established business lines and businesses that depend more heavily on renting out AI computing capacity.

SpaceX has satellite internet and launch operations alongside its AI ambitions.

Those businesses may provide additional sources of revenue, although they do not guarantee that the AI infrastructure investment will deliver its projected returns.

The transaction alone does not prove that smaller AI cloud providers will fail or that the market will consolidate.

It does show that access to capital, the cost of advanced hardware, and the ability to generate revenue from computing infrastructure are important competitive factors.

What Happens Next

SpaceX expects the master finance lease to close before the end of 2026, subject to specified conditions.

Lease schedules are expected to commence as GPUs and related equipment are delivered during the first half of 2027.

Several developments will determine how the arrangement progresses: whether the definitive lease closes on the expected timeline, when the equipment arrives, what financing rate is ultimately agreed, and how quickly SpaceX deploys the hardware at its selected data centers.

The final lease terms will also matter for assessing the total cost of the financing.

SpaceX has disclosed the expected aggregate principal payments, but interest and other expenses add to the overall financial commitment.

For businesses watching the AI infrastructure market, the central takeaway is that access to computing hardware increasingly depends not only on technology suppliers but also on sophisticated financing arrangements.

SpaceX’s agreement with an Apollo-managed entity gives the company a structured path to acquire next-generation AI equipment while spreading payments across multiple years.

Its long-term value will depend on execution, computing demand, and the company’s ability to turn that infrastructure into sustainable revenue.

Why is SpaceX using a leasing arrangement instead of directly purchasing all the hardware?

The structure allows an Apollo-managed entity to acquire the equipment and lease it to SpaceX, spreading payments across the lease periods rather than requiring SpaceX to fund the entire acquisition directly at the outset.

The arrangement still creates substantial future payment obligations, and its overall cost will depend on interest and other expenses.

Key Facts

• Announced: October 26, 2026.

• Financing structure: A master finance lease with an Apollo Global Management-managed entity.

• Expected aggregate principal payments: Approximately $40 billion, excluding interest and certain additional costs.

• Planned hardware: NVIDIA Vera Rubin GPUs and related equipment.

• Expected timeline: Closing before year-end 2026, with equipment deliveries and lease schedules expected to begin in the first half of 2027.

Editorial Notes

The financing commitment is binding, but the master lease remains subject to specified closing conditions.

The final interest rate and certain commercial terms had not been disclosed in the public announcement.

The $40 billion figure represents expected aggregate principal payments over the lease schedules, not an immediate cash payment to SpaceX.

Forecasts regarding deployment, revenue, and investment returns remain subject to execution and market risks.