Nvidia Bets $40B on Its Own AI Customers

Nvidia has crossed $40 billion in equity commitments this year, and the pattern is difficult to ignore. The company is investing in its own customers - data center operators, cloud providers, and AI labs - who then use that capital to buy more Nvidia chips. The result is a self-reinforcing loop that has turned the world's most valuable chip company into one of AI's largest venture investors in under twelve months.

The Investment Ledger

The headline number is $30 billion, which Nvidia put into OpenAI in February after its proposed $100 billion infrastructure partnership with the AI lab quietly collapsed. That single equity stake passes what most sovereign wealth funds deploy in a year. But the supporting deals reveal the full architecture of what Nvidia is building as an investor.

The Corning deal is the most unusual of the group. Nvidia committed $500 million right away - through warrants giving it the right to buy shares - with the option to invest up to $3.2 billion total. In exchange, Corning is building three new advanced manufacturing facilities in North Carolina and Texas, dedicated completely to optical fiber technologies for Nvidia's rack-scale systems. The deal also creates more than 3,000 jobs and will increase Corning's US-based fiber production capacity by more than 50%.

Jensen Huang has been the architect of Nvidia's push from chipmaker to AI ecosystem investor. Source: commons.wikimedia.org

The IREN deal is the clearest illustration of how these arrangements work. Nvidia invested $2.1 billion in the data center operator, and IREN simultaneously committed to purchasing $3.4 billion in cloud services from Nvidia over five years. Equity flows one direction; revenue flows the other.

Beyond the public deals, Nvidia participated in roughly two dozen private startup rounds during 2026 alone, following 67 venture deals across 2025. The company isn't just selling chips anymore. It's financing the companies that run on them.

Who Benefits

Nvidia's demand insurance

Nvidia's core business thesis is that AI infrastructure will produce at least $1 trillion in chip revenue from its Blackwell and Rubin GPU lines through the end of 2027 - a figure Huang cited at GTC 2026 in March. The equity investments function as demand insurance. When Nvidia puts $2 billion into Nebius - an AI cloud provider that deploys full-stack Nvidia infrastructure - it's securing a customer that is structurally committed to buying more of what Nvidia makes.

Nebius already had that commitment before the investment. After the deal closed in March, the company landed a five-year, $12 billion infrastructure agreement with Meta, then a separate multi-billion-dollar deal with Microsoft. Its stock has jumped 57% since the Nvidia investment was announced. Nvidia's check opened doors, and the GPU purchases followed.

Companies receiving capital

For the AI startups and infrastructure operators on the other side of these deals, Nvidia's money comes with advantages beyond the capital itself. Early access to new chip generations. Preferred partnership status. The signal that the industry's most important hardware company has staked its credibility on your growth. These benefits are difficult to price but easy to observe - the companies Nvidia invests in tend to close more deals afterward, often with each other.

"Our investments are focused very squarely, strategically on expanding and deepening our ecosystem reach," Huang said in his comments on the investment strategy.

Jensen Huang's information advantage

Huang's vantage point as the operator of the most critical hardware company in the AI supply chain gives him access to data no fund manager can replicate. Every deal he evaluates reflects real-time information about which companies are actually scaling compute, not just announcing it. He knows which hyperscalers are ordering next-generation racks before the purchases are disclosed in filings. The equity portfolio is capital deployment and market intelligence at the same time.

Nvidia's equity investments lock customers into infrastructure built on its accelerators. Source: unsplash.com

Who Pays

Nvidia shareholders carry the circular risk

Matthew Bryson at Wedbush Securities put the concern plainly: Nvidia's dealmaking fits "squarely into the circular investment theme," referring to the pattern where a company invests in its own customers, who then use those funds to buy from the same company. If Nvidia's chip demand softens - due to a breakthrough that cuts compute requirements, a competitor closing the performance gap, or a broad economic contraction - the equity positions in its own customers won't offer diversification. They'll amplify the exposure.

The company's record revenue and stock performance have muted this criticism so far. But the strategy is only tested when demand softens, and no one can predict when that happens.

Competing chipmakers squeezed out

The less visible cost of Nvidia's equity strategy lands on AMD, Intel, and anyone else trying to win data center business from the same companies Nvidia has invested in. When Nvidia becomes a shareholder in CoreWeave and IREN, it's no longer just a supplier to those companies - it's an equity partner whose interests are aligned with theirs. The commercial relationship and the capital relationship become the same relationship.

Nvidia's capacity to offer this kind of capital is something AMD and Intel can't match at current scale. That's a structural advantage that has nothing to do with transistor counts, and it didn't exist two years ago.

The question of returns

The $30 billion OpenAI stake is Nvidia's largest single bet, and it comes with no guaranteed timeline for liquidity. OpenAI's most recent funding round pegged its valuation at around $850 billion. At that price, Nvidia paid roughly $30 billion for a stake at a valuation where the margin for error is limited. If OpenAI's implied value softens before an IPO, the equity position becomes expensive regardless of how much GPU revenue it anchors.

Bryson's note that successful investments could help Nvidia build a "competitive moat" is the optimistic read. The pessimistic read is that Nvidia has written $40 billion in checks to customers who might eventually find other chip suppliers - and taken those positions at peak-cycle AI valuations.

Nvidia has always profited from the AI boom by selling picks and shovels. The equity strategy is a bet that it can profit from the boom twice - once when customers buy the chips, and again when those customers eventually go public.

Sources:

• Nvidia has already committed $40B to equity AI deals this year - TechCrunch
• NVIDIA and Corning Announce Long-Term Partnership - NVIDIA Newsroom
• NVIDIA and Nebius Partner to Scale Full-Stack AI Cloud - NVIDIA Newsroom
• NVIDIA and IREN Announce Strategic Partnership - NVIDIA Newsroom
• NVIDIA and IREN Partnership Details - GlobeNewswire