SpaceX Seeks Staggering $40 Billion To Buy Nvidia AI Chips In Massive Compute Gamble

Written By James Taylor
Musk’s SpaceX Wants $40 Billion For Nvidia Chips As AI Infrastructure Race Accelerates

SpaceX Bets Big On Nvidia

Musk’s Giant Nvidia Spending Plan

The reported $40 billion financing would rank among the clearest signs yet that the AI race is becoming a contest over who can afford the largest computing infrastructure.

SpaceX is reportedly seeking around $40 billion in new financing to fund the purchase of Nvidia artificial intelligence chips, in what would be one of the most aggressive hardware bets of the current AI boom.

The proposed structure is striking. Around $10 billion would come from bank loans, while roughly $30 billion would be raised through investment-grade debt. Apollo Global Management is expected to lead the financing effort and help place the debt with investors, while Pimco is among the institutions reported to have held talks about participating.

The transaction remains under discussion. It has not been announced as a completed financing package, and the companies involved had not publicly confirmed the full reported structure when the story emerged.

If it goes ahead, however, the deal would reveal the scale of SpaceX’s ambitions far beyond rockets and satellites.

Why SpaceX Wants So Much Computing Power

SpaceX is no longer simply a launch company.

Its artificial intelligence operations have become increasingly important after the integration of xAI, the business behind Grok and the Colossus computing infrastructure. That combination has created an unusual corporate structure spanning rockets, satellites, communications, AI models and enormous data-centre requirements.

The company has already committed to Nvidia’s latest hardware. Nvidia said in August that SpaceXAI would deploy Vera CPUs and expand its AI infrastructure around the Vera Rubin platform as it scaled towards gigawatts of computing capacity.

That matters because the most advanced AI systems are hungry for far more than clever software. They require huge clusters of specialised processors, high-speed networking, cooling systems, storage, electricity and buildings capable of supporting extraordinary amounts of heat and power demand.

Nvidia sits near the centre of that supply chain.

Its Vera Rubin architecture is designed for the next generation of large-scale AI systems, bringing together GPUs, CPUs, networking and other components into tightly integrated computing platforms. Nvidia says the technology is intended to improve the cost and efficiency of training and running increasingly complex AI models.

SpaceX’s reported financing suggests it does not want a modest allocation of that technology. It wants scale.

The $40 Billion Structure Explained

The proposed financing is unusual not simply because of its size, but because of how it would be funded.

About one quarter of the total would reportedly come from bank loans. The remaining three quarters would come from investment-grade debt sold to institutional investors.

That could include asset managers, pension funds, insurers and other large investors able to buy high-grade corporate bonds.

Apollo’s reported role is important. A financing of this size would need to be structured, marketed and distributed across a broad investor base. A single lender is unlikely to carry anything close to the entire exposure.

SpaceX’s financial position has also changed significantly in 2026.

The company completed its initial public offering in June, raising approximately $85.7 billion in gross proceeds. Its shares began trading under the ticker SPCX, giving the company access to public equity markets as well as a potentially deeper pool of debt financing.

That IPO provides important context. The reported $40 billion package would not be the rescue financing of a cash-starved business. It would represent an attempt to use the balance sheet of a newly public technology giant to accelerate spending on AI infrastructure.

The distinction matters.

Why Nvidia Chips Have Become Strategic Assets

The AI industry once competed mainly through talent, algorithms and access to data.

It now competes just as aggressively through physical infrastructure.

Training frontier models requires thousands, and increasingly tens of thousands, of high-end accelerators operating together. Serving those models to millions of users creates another enormous computing requirement.

That has turned advanced processors into strategic assets.

Companies with reliable access to the newest chips can train more ambitious models, run larger experiments and support more users. Companies without enough hardware can be forced to slow development, rent capacity from rivals or wait for scarce infrastructure to become available.

The result is a race not only to design better AI, but to secure the industrial machinery behind it.

That broader shift is already visible across the technology sector. AI developers are signing long-term infrastructure agreements, constructing gigantic data centres and exploring increasingly creative forms of financing to pay for them.

The cost is beginning to resemble heavy industry.

SpaceX And Nvidia Are Already Deeply Linked

The proposed financing would deepen an existing relationship.

Nvidia announced in August that SpaceXAI would use Vera CPUs and expand around its Vera Rubin platform. The companies also described plans to extend Nvidia computing technology into space through a first-generation Starmind AI satellite.

That makes the relationship broader than an ordinary supplier contract.

SpaceX has the launch infrastructure. Nvidia has the processors. SpaceX also operates Starlink, giving it a global communications network that could eventually interact with AI services distributed across terrestrial and orbital infrastructure.

The combination points towards something more ambitious than a conventional data-centre buildout.

The growing political and technological influence surrounding Elon Musk has often focused on Tesla, SpaceX and X as separate businesses. Increasingly, however, Musk’s companies are becoming linked through AI, communications, engineering and computing.

That convergence could become one of the defining features of his corporate strategy.

The Risk Behind The Spending

A $40 billion chip financing would also carry significant risk.

AI hardware develops quickly. A processor bought at enormous cost can remain useful for years, but the market value of cutting-edge chips can fall rapidly when newer generations arrive.

Lenders therefore face a difficult question: how should they value hardware that may generate huge amounts of revenue but could also depreciate far faster than traditional industrial assets?

That issue has become increasingly important as banks and investment funds are asked to finance more AI infrastructure.

There is also the question of utilisation.

The economics of an AI data centre depend heavily on keeping expensive chips busy. Hardware that sits idle still consumes capital, space and maintenance costs. The financial case therefore depends on sustained demand for AI training and inference.

SpaceX appears confident that demand will remain strong.

Its own Grok systems need computing capacity, while the wider market for rented AI infrastructure continues to expand. The company’s vertically integrated position could also give it options unavailable to a standalone data-centre operator.

Still, the sums involved are exceptional.

AI Is Becoming A Capital Markets Story

The most important part of the reported deal may not be SpaceX itself.

It is what the financing says about the AI industry.

Artificial intelligence began its current boom as a software story. It is increasingly becoming an energy, construction and capital-markets story.

The companies attempting to build the most advanced systems now need funding on a scale once associated with telecommunications networks, energy infrastructure and national industrial projects.

That change will affect who can compete.

A small team can still develop an important model or application. But building infrastructure at the frontier increasingly requires access to billions of dollars, enormous power supplies and relationships with semiconductor manufacturers.

The rapid expansion of AI into everyday commercial systems is already creating new legal and economic questions. At the infrastructure level, an even bigger shift is under way: AI is becoming one of the most capital-intensive technology industries ever created.

What Happens Next

The reported financing is expected to progress towards a possible closing in 2027, although discussions remain at an early enough stage that the final size, structure or participants could change.

That uncertainty should not be overlooked.

The headline number is enormous, but it is not yet money in the bank or chips installed in a data centre.

What is clear is the direction of travel.

SpaceX has raised unprecedented amounts of equity, integrated AI more deeply into its business, committed to Nvidia’s latest computing architecture and is now reportedly exploring tens of billions of dollars in additional financing.

The Pentagon’s decision to bring Musk into long-term technology planning also shows how his influence increasingly reaches across AI, space, communications and national security.

If the $40 billion Nvidia financing closes, it will not merely be a large chip purchase.

It will be another step towards a world in which the most powerful AI systems are built by companies capable of financing infrastructure on the scale of major industrial economies.

Sources

Next Reads

James Taylor
Next

McDonald’s Sued Over Alleged AI Price-Fixing As Big Mac Pricing System Faces Antitrust Test