Nvidia Is Betting $40 Billion on the AI Companies That Buy Its Chips

Nvidia has already committed more than $40 billion to equity investments in AI companies in the first five months of 2026 alone and the biggest single check, a jaw-dropping $30 billion, went to OpenAI, the company that depends on Nvidia GPUs to power the very products that made it famous. The question the AI industry is now asking: is this visionary empire-building, or is it one very elaborate loop?

Key Takeaways

• Nvidia crossed $40 billion in AI equity commitments in 2026 most in just five months
• A single $30 billion bet on OpenAI accounts for 75% of that total
• Seven additional multi-billion investments target companies across the AI supply chain
• Critics flag a “circular investment” concern Nvidia funds companies that buy Nvidia chips
• Jensen Huang calls it strategic ecosystem deepening; Jensen’s bet on Intel has already returned 5x
• Nvidia generated $97 billion in free cash flow last fiscal year it has the firepower to do this

From Chipmaker to Ecosystem Bank

For most of its existence, Nvidia was the company that made the hardware everyone else needed. It sold GPUs. It watched others build on top of them. That’s the boring version of Nvidia. The 2026 version is something different.

According to reporting by TechCrunch and CNBC, Nvidia has now committed more than $40 billion to equity investments in AI companies all in the first five months of 2026. That’s a pace that makes even the most aggressive venture funds look cautious. The centerpiece of it all is a $30 billion equity stake in OpenAI, which alone represents three-quarters of the total capital deployed this year.

Then there’s the wider portfolio. Nvidia has announced seven separate multi-billion-dollar investments in publicly traded companies, including up to $3.2 billion in Corning the glassmaker that produces optical components critical for AI data center interconnects and up to $2.1 billion in IREN, a data center operator actively expanding capacity for AI workloads. The chip company has also joined roughly two dozen funding rounds for private AI startups, with additional stakes in Marvell, Lumentum, Coherent, CoreWeave, and Nebius. This isn’t a chip company dabbling in venture. This is a chip company systematically buying pieces of the entire food chain it feeds.

The Investment That Started It All and Already Paid Off

To understand why Jensen Huang is moving this aggressively, it helps to look at what already happened. In 2025, Nvidia placed a $5 billion bet on Intel. That bet is now worth over $25 billion a 5x return in a matter of months, at a time when Intel stock has surged more than 200%. That’s not luck. That’s what happens when the company that powers the AI boom also happens to invest in the company making the hardware that AI boom increasingly demands.

The Intel win gave Nvidia’s leadership both confidence and a playbook. If you invest in the companies your technology enables, and those companies succeed because of your technology, the returns compound in two directions simultaneously on the investment and through continued product sales. It’s a self-reinforcing flywheel, and Nvidia is now spinning it at full speed.

Nvidia’s own balance sheet backs this up. The company generated $97 billion in free cash flow in its last fiscal year. Its non-marketable equity securities (private company investments) swelled from $3.39 billion to $22.25 billion in just one year, while gains on public and private investments hit $8.92 billion up from $1.03 billion the prior year. This is no longer a side project. Investment returns are becoming a meaningful line item in Nvidia’s financials.

Why the Corning and IREN Deals Are More Interesting Than They Look

Most coverage of Nvidia’s investment spree fixates on the OpenAI number because $30 billion is almost comically large. But the Corning and IREN deals reveal something more telling about Nvidia’s actual strategy.

Corning doesn’t make AI models. It makes glass. Specifically, it produces specialized optical fiber and components used in the physical infrastructure that connects AI data centers. As GPU clusters scale to tens of thousands of chips, the bottleneck increasingly isn’t compute it’s interconnect bandwidth. Getting data in and out of these clusters fast enough is now an engineering challenge at the physical layer. Nvidia investing in Corning is Nvidia saying it sees that bottleneck and wants to own a piece of the solution before competitors can.

IREN, meanwhile, is a data center operator expanding specifically for AI workloads. By taking a stake in IREN, Nvidia gains alignment with a company that will be ordering significant amounts of Nvidia hardware over the coming years. These aren’t passive financial bets. They’re strategic anchor investments designed to pull demand through the supply chain.

This is also consistent with how we’ve seen other tech giants operate as we reported, Jeff Bezos raised $100 billion for an AI manufacturing fund with a similar logic of controlling physical infrastructure rather than just software. The pattern across the AI industry is clear: the real leverage is at the layer nobody’s watching.

The “Circular Investment” Criticism And Why It Deserves Serious Attention

Not everyone is impressed. Matthew Bryson, an analyst at Wedbush Securities, described Nvidia’s dealmaking as fitting “squarely into the circular investment theme” a phrase that captures the core concern succinctly. The worry goes like this: Nvidia invests in OpenAI. OpenAI uses that capital to buy Nvidia GPUs. Nvidia records revenue. The investment looks like independent demand, but it’s actually Nvidia paying itself through a middleman.

Some analysts have gone further, comparing the dynamic to vendor financing schemes that helped inflate the dot-com bubble where companies lent money to customers so those customers could buy their products, making sales figures look healthier than underlying demand actually supported.

Bryson himself offered a more nuanced take: the investments could help create a “competitive moat” if Nvidia executes well. And that’s probably the right frame. Circular investment concerns are real, but they’re most damaging when the underlying products have no genuine demand. Nvidia’s GPUs remain supply-constrained in ways that suggest real, independent demand not manufactured demand from recycled investment capital.

How This Compares to What Google, Microsoft, and Amazon Are Doing

Nvidia isn’t operating in a vacuum. Microsoft invested $13 billion in OpenAI over several years before Nvidia wrote its $30 billion check. Google and Amazon have poured billions into Anthropic. The hyperscalers have been doing strategic AI investing for half a decade. What makes Nvidia’s position distinctive is the nature of the dependency it creates.

When Microsoft invests in OpenAI, OpenAI’s products run on Microsoft’s Azure cloud. But OpenAI can theoretically switch cloud providers. Switching GPU architectures after training models on CUDA, after optimizing inference pipelines around Nvidia hardware, after building entire engineering cultures around Nvidia tooling is an entirely different level of switching cost. Nvidia’s investments don’t just create financial alignment. They reinforce an ecosystem lock-in that makes the dependency structural rather than contractual.

That’s the competitive moat Bryson was pointing to. And it’s much harder to dismantle than a cloud contract.

The broader shift is worth tracking too. As companies that replaced workers with AI agents are now reconsidering that strategy, the demand profile for AI infrastructure is getting more complex and Nvidia’s investments are essentially a hedge against any particular use case failing, since they’re spread across the infrastructure stack rather than any single application layer.

The OpenAI Bet: Why $30 Billion Is Not Just a Number

The $30 billion OpenAI investment deserves separate analysis because of its sheer scale. No single corporate AI investment in history comes close to this figure. Microsoft’s multi-year commitment to OpenAI totaled around $13 billion. Nvidia’s single transaction is more than double that, made in what appears to be a single tranche in early 2026.

There’s a reported backstory here worth noting. CNBC had previously written about discussions around a $100 billion partnership between Nvidia and OpenAI that was eventually restructured into this $30 billion equity stake. That gap from $100 billion to $30 billion suggests the terms evolved significantly during negotiations. The final number reflects real due diligence about OpenAI’s valuation and revenue trajectory, not just an uncritical bet on hype.

For context, OpenAI is currently building a desktop superapp that could consolidate all its products into a single interface a move that would significantly increase the scale of compute it needs. If that product succeeds, Nvidia’s $30 billion bet looks prescient. If OpenAI’s growth hits a ceiling, the bet looks overpriced. Either way, Nvidia now has a seat at the table when OpenAI makes decisions that affect its infrastructure choices.

What This Strategy Means for the Rest of the AI Industry

Nvidia’s moves have implications that extend well beyond its own balance sheet. If the dominant AI chip supplier is also a major equity holder in the leading AI model companies, the competitive landscape gets complicated in interesting ways.

For AI startups seeking to compete with OpenAI, Anthropic, or Google DeepMind: Nvidia is increasingly financially aligned with the incumbents. That doesn’t mean Nvidia will deny chips to competitors its business model depends on broad adoption but it does mean that the relationship between Nvidia and a well-capitalized incumbent goes beyond a vendor-customer dynamic. There’s shared equity, shared incentives, and shared risk.

For investors watching the AI sector: the circular investment concern is a real risk factor that deserves a place in any valuation model. If AI company revenues are partly dependent on capital recycled from chip company investments, the true organic demand for AI products is harder to measure than the headline numbers suggest.

And for developers and enterprises building on top of this infrastructure: Nvidia’s vertical integration push is a signal that the hardware layer of AI is no longer a neutral utility. The company selling you the GPUs also owns a piece of the models running on them. That’s a different competitive environment than the one that existed two years ago. Understanding how developers use AI tools in this new environment will increasingly mean understanding who owns what in the stack below them.

What Happens Next

Nvidia’s $40 billion in commitments is almost certainly not the end. The company has the cash flow to sustain this pace, the track record to justify it internally, and the competitive rationale to keep going. Jensen Huang has spent the last decade turning a graphics card company into the backbone of the global AI industry. The investment strategy of 2026 looks like phase two of that same vision not just supplying the picks and shovels to the gold rush, but owning stakes in the mines themselves.

The questions that remain whether this creates a regulatory problem, whether the circular investment concerns are overblown or underappreciated, whether OpenAI delivers the kind of returns that justify a $30 billion bet won’t be answered this year. But they will be answered. And the answers will shape what the AI hardware and software market looks like for the next decade.

In the meantime, one thing is certain: Nvidia is no longer just the company that makes the chips. It’s the company that owns the companies that use the chips. That’s a fundamentally different kind of power and it’s worth paying close attention to.