- Nvidia has committed more than $40 billion to equity AI investments in the first months of 2026, topping its entire 2025 pace
- The chipmaker’s $5 billion Intel stake has already quintupled to $25 billion, but its neocloud bets face growing ‘circular financing’ criticism
- Short seller Jim Chanos compares Nvidia’s strategy to Lucent’s vendor financing before the dot-com crash
Nvidia doesn’t just sell GPUs anymore. It bankrolls the companies that buy them.
The world’s most valuable company has committed more than $40 billion to equity investments in AI companies in 2026 alone, according to TechCrunch. That figure, confirmed by CNBC, covers just the first few months of the year and dwarfs the $17.5 billion Nvidia invested across all of last fiscal year.
Most of that $40 billion is a single check: the $30 billion Nvidia pumped into OpenAI in February. But the long tail of deals tells a more interesting story. This week alone, Nvidia signed agreements to invest up to $3.2 billion in Corning and up to $2.1 billion in data center operator IREN. In March, it put $2 billion each into Marvell Technology, Lumentum, and Coherent. In January, another $2 billion into CoreWeave. The pattern is unmistakable: Nvidia is financing the entire AI supply chain, from glass to photons to compute.
The $25 Billion Intel Trade Nobody Talks About
Lost in the headline numbers is Nvidia’s single best trade. Its $5 billion investment in Intel, negotiated at $23 per share, has already quintupled. Intel stock has surged past 200% gains this year, making Nvidia’s position worth more than $25 billion, as The Register reported. That is a historic return in a matter of months, and it came from bailing out a struggling US chipmaker that also happens to be a potential foundry partner.
On Nvidia’s balance sheet, the numbers tell the story of an increasingly financial operation. Non-marketable equity securities swelled from $3.39 billion in January 2025 to $22.25 billion a year later. Realized gains on equity assets jumped from $1.03 billion to $8.92 billion in the same period, per CNBC. Nvidia generated $97 billion in free cash flow last fiscal year. It has capital to deploy, and it is deploying it aggressively.
‘It Smells Like You Are Pre-Funding Your Own GPUs’
The deals break into two categories, and analysts are split on which one matters more. Component-maker investments — Corning for fiber optics, Marvell and Lumentum for silicon photonics — draw praise. Jordan Klein at Mizuho called them “super smart by the CFO and team and a great use of cash” in a note reported by CNBC, because they accelerate development of technology Nvidia actually needs.
The neocloud bets are a different story. When Nvidia puts $2 billion into CoreWeave and CoreWeave uses that money to build data centers powered by Nvidia chips, the circularity is hard to ignore. Klein was blunt: “It smells like you are pre-funding the purchase of your own GPUs and products.” Ben Bajarin at Creative Strategies put it differently, telling CNBC that “the risk is that if the cycle turns, the market starts questioning how much of the demand was organic versus supported by Nvidia’s own balance sheet.”
Matthew Bryson, an analyst at Wedbush Securities, said Nvidia’s investments fit “squarely into the circular investment theme” that has fueled fears about the durability of AI spending. But Bryson also acknowledged that if Nvidia executes, the deals could build a “competitive moat.”
Michael Cusumano, a professor at MIT Sloan, was more direct about the OpenAI deal. Per the TechCrunch report citing the Financial Times, he said the original $100 billion commitment was “kind of a wash,” since Nvidia was investing that money into OpenAI while OpenAI was simultaneously committing to buy $100 billion or more of Nvidia chips.
The Lucent Parallel
For anyone who watched the dot-com bubble, Nvidia’s strategy has an uncomfortable precedent. Lucent Technologies, the telecom equipment giant, extended loans and financing to loss-making startup customers in the late 1990s. Those startups bought Lucent equipment with Lucent’s own money. When the dot-com crash wiped out those customers, Lucent wrote off billions in uncollectible loans and never recovered.
Jim Chanos, the short seller who famously predicted Enron’s collapse, sees the parallel. As reported by the Economic Times, Chanos said Nvidia is “putting money into money-losing companies in order for those companies to order their chips.” The comparison is not exact — Nvidia is buying equity, not extending loans, and its chips are physical assets with real demand. But the structural resemblance is real enough that Nvidia felt compelled to respond.
In November, the company sent a seven-page memo to Wall Street analysts denying it was engaged in vendor financing. The memo, first reported by Barron’s, argued that Nvidia’s business is “economically sound” and its reporting “complete and transparent.” Jensen Huang has called the circular financing label “ridiculous,” telling Bloomberg that Nvidia’s investments are “a small percentage of the overall amount of infrastructure capital” its partners need to raise.
There is a key difference between Nvidia and Lucent that the company’s defenders point out: demand. Nvidia’s GPUs have a months-long backlog. OpenAI, Anthropic, xAI — they all need Nvidia chips regardless of who funded the data center. As Frontierbeat reported, even SpaceX is now in the compute supply business. The question is not whether the chips are useful. It is whether the investments are creating demand that would not otherwise exist.
Huang Is Already Pulling Back
The most telling signal may be that Nvidia itself appears to be winding down the strategy. Huang said at the Morgan Stanley conference in March that the $30 billion OpenAI check “might be the last time” Nvidia invests before the company’s anticipated IPO. He offered a straightforward explanation: once OpenAI and Anthropic go public, the pre-IPO investment window closes.
That is plausible. It also happens to be convenient. The original OpenAI commitment was $100 billion. It shrank to $30 billion. Huang’s explanation for the pullback focused on opportunity cost, but the shrinking commitment coincided with mounting criticism of the circular financing model. Nvidia’s relationship with Anthropic has also been complicated — the AI lab’s CEO, Dario Amodei, compared US chip sales to approved Chinese buyers to “selling nuclear weapons to North Korea” at Davos, weeks after Nvidia wrote a $10 billion check. The Trump administration later blacklisted Anthropic from federal contracts.
For now, Nvidia’s $5.2 trillion valuation and $97 billion cash flow give it plenty of room to keep placing bets. The company booked $8.92 billion in investment gains last year on top of its chip revenue. But as Frontierbeat noted when OpenAI missed its revenue targets, the math on AI compute spending only works if the demand is real. Nvidia’s $40 billion investment spree is a bet that it is. If the cycle turns, that bet becomes a balance sheet problem.
Nvidia reports fiscal first-quarter earnings in less than two weeks. Shareholders will get their clearest look yet at the size of the investment portfolio — and how much of the chipmaker’s growth is coming from customers it paid to exist.
FAQ
What is circular financing or vendor financing?
Vendor financing is when a company invests in or extends loans to its own customers, who then use that money to buy the company’s products. It inflates reported revenue but creates risk if the customer can’t sustain purchases without the financing. Lucent Technologies famously used this model before the dot-com crash.
How much has Nvidia invested in AI companies in 2026?
Nvidia has committed more than $40 billion to equity AI investments in 2026, per CNBC and TechCrunch. The largest single investment is $30 billion in OpenAI, with additional multibillion-dollar deals in Corning, IREN, CoreWeave, Nebius, Marvell, Lumentum, and Coherent.
What was Nvidia’s return on the Intel investment?
Nvidia invested $5 billion in Intel at $23 per share. With Intel stock up over 200% in 2026, that position is now worth more than $25 billion — a 5x return in a matter of months.
Why are short sellers comparing Nvidia to Lucent?
Jim Chanos, the short seller who predicted Enron’s collapse, says Nvidia is “putting money into money-losing companies in order for those companies to order their chips” — the same model Lucent used with telecom startups in the late 1990s. When those startups failed, Lucent lost billions.
Is Jensen Huang pulling back from AI investments?
Huang said at the March 2026 Morgan Stanley conference that the $30 billion OpenAI investment “might be the last time” Nvidia invests before OpenAI’s IPO. He called the circular financing criticism “ridiculous” and said Nvidia’s stakes are a small fraction of what partners need to raise.
