Nvidia’s $40B AI equity spree raises new questions

Nvidia has topped $40 billion in 2026 equity commitments across the AI supply chain.

Nvidia is no longer just selling GPUs. In 2026, the company has pushed past $40 billion in equity commitments, with fresh deals for IREN and Corning adding $5.3 billion of potential investment in a single week.

That matters because Nvidia is using its cash pile to shape the market that buys its own hardware. The company reported $97 billion in free cash flow last fiscal year, and its stock has climbed more than 11-fold in four years, pushing its market value to about $5.2 trillion.

Why Nvidia is writing checks all over AI

Nvidia’s strategy is simple to describe and hard to ignore: put money into the companies that build, power, cool, and buy AI systems, then sell them more Nvidia hardware. That includes model makers like OpenAI, cloud infrastructure firms such as CoreWeave, and newer infrastructure players like Nebius.

The logic is straightforward. AI demand is still constrained by chips, power, networking, and data center space. If Nvidia helps fund the companies that relieve those bottlenecks, it also helps expand the market for its own systems. That is why the company is not limiting itself to one corner of the stack.

• IREN gives Nvidia exposure to data center buildout.
• Corning links Nvidia to fiber and optical components.
• OpenAI, CoreWeave, and Nebius tie Nvidia to compute demand.
• Marvell, Lumentum, and Coherent connect Nvidia to photonics and silicon work.

The pattern is bigger than a few headline deals. Nvidia has signed at least seven multibillion-dollar investments with public companies this year and has taken part in roughly two dozen private rounds, according to FactSet. That is a lot of capital, and it is moving fast.

The circular money problem is getting louder

There is a real debate here, and it is not subtle. Some investors see smart supply-chain planning. Others see a loop where Nvidia funds the same companies that then buy Nvidia gear, which makes demand look stronger than it might be on its own.

Matthew Bryson of Wedbush Securities said Nvidia’s dealmaking fits “squarely into the circular investment theme.” That is the cleanest way to describe the worry: money goes out, orders come back, and the line between organic demand and financed demand gets blurry.

“It smells like you are pre-funding the purchase of your own GPUs and products,” said Jordan Klein, chip analyst at Mizuho.

That criticism matters because Nvidia’s balance sheet is now part of the story. The company said its non-marketable equity securities rose to $22.25 billion at the end of January, up from $3.39 billion a year earlier. It also reported $8.92 billion in gains on private and public holdings, compared with $1.03 billion in the prior fiscal year.

Those are not side notes. They are proof that Nvidia’s investments are already affecting earnings, valuation, and investor expectations. If the stock market starts to wobble, those gains can shrink fast.

How the biggest bets compare

The scale of Nvidia’s portfolio is easier to understand when you line up the numbers. The company’s $30 billion stake in OpenAI is still the headline bet, but the Intel position is the one that has produced the most striking mark-to-market gain so far.

Intel’s stock is up well over 200% this year, which pushed Nvidia’s $5 billion investment there to more than $25 billion in value. That is a huge paper gain, but it also shows how much Nvidia’s results now depend on equity markets, not just chip shipments.

• Intel: Nvidia’s $5 billion bet is now worth over $25 billion.
• Lumentum and Coherent: each received a $2 billion Nvidia investment in March.
• Marvell Technology: got a separate $2 billion strategic investment.
• CNBC’s report says Nvidia’s 2026 commitments already exceed $40 billion.

The IREN and Corning deals also point to a broader shift in what Nvidia wants to own around the GPU business. IREN is expected to deploy up to 5 gigawatts of Nvidia’s DSX-branded infrastructure designs, while Corning is building three new U.S. facilities for optical technologies tied to Nvidia’s rack-scale systems.

That suggests Nvidia is not just financing demand. It is shaping the physical infrastructure that demand depends on, from glass and fiber to power and rack design.

What this means for the next earnings call

Nvidia’s next earnings report should tell investors whether this strategy is still a smart use of cash or whether it is starting to distort the picture. The company has already said its investments are meant to “expand and deepen” its ecosystem, and Jensen Huang has argued that Nvidia does not pick winners, it supports the whole field.

The problem is that markets rarely care about intent when the numbers get this large. If more of Nvidia’s growth story is tied to equity stakes, the company will have to prove that demand for GPUs remains strong even without financial support attached.

My read is simple: watch the mix of revenue growth, investment gains, and customer concentration in the next report. If Nvidia keeps turning cash into supply-chain control without slowing chip sales, rivals will have a harder time catching up. If the loop starts to look self-funded, the questions around AI demand will get much louder.

For now, the headline is clear: Nvidia is acting like an industrial investor as much as a chip company, and the market has to decide whether that is smart capital allocation or a sign that AI infrastructure still needs a financial backstop.