Nvidia's $30B OpenAI Bet May Be Its Last

When Jensen Huang stepped up to the microphone Wednesday, his words carried more weight than usual. The Nvidia CEO announced a $30 billion investment in OpenAI but added a crucial caveat: "This might be the last time" the chip giant bets on Sam Altman's AI darling before it goes public.

That qualifier sent ripples through Silicon Valley. After $168 billion in total funding, OpenAI remains the most valuable startup in history—yet it still lacks a proven path to profitability. Now, even its biggest backers are starting to hedge their bets.

The Math Behind the Hesitation

Nvidia's $30 billion commitment isn't pocket change, even for the world's most valuable company. It represents an eighth of Nvidia's annual revenue and half of its just-reported quarterly earnings of $68 billion—a figure that beat expectations by 73% year-over-year.

Yet Nvidia's stock tumbled 9% that week despite the stellar earnings. The message from investors was clear: we're not convinced these massive AI bets will pay off.

"I don't think anyone knows how to properly value anything surrounding AI," explains Aleksandar Tomic, associate dean at Boston College. "We're still waiting to see how these companies will monetize what they produce. The potential is enormous, but it's like the internet in the late '90s—the promise is there, even if the business model isn't fully formed yet."

The Trillion-Dollar Question

The numbers tell a sobering story. HSBC initially forecast that OpenAI's computing obligations would reach $1.4 trillion by 2033. OpenAI pushed back, revising the estimate to $600 billion by 2030, but even that "modest" figure includes $620 billion just for data center rental space.

To justify its current $730 billion valuation, OpenAI needs to generate $200 billion in annual revenue by 2030—a 15x increase in five years while costs explode exponentially. The company currently serves 900 million users, but most aren't paying customers.

Microsoft, OpenAI's largest backer with $13.75 billion invested since 2019, felt similar investor skepticism. Despite strong earnings in January, Microsoft's stock dropped 11% when investors spotted slowing growth in its Azure cloud platform—the very service that hosts OpenAI's technology.

The Circular Investment Web

What's particularly concerning to analysts is the circular nature of these investments. Nvidia invests in OpenAI, which then commits to buying Nvidia chips. Microsoft funds OpenAI, which signs a $250 billion contract for Microsoft's Azure services. Oracle agreed to a $300 billion deal to build data centers that OpenAI will then rent back.

"A lot of circular deals, right?" notes Tomic. "That's reminiscent of the early 2000s" dotcom bubble.

This pattern extends beyond the core players. Disney invested $1 billion in December for access to OpenAI's video generation platform Sora, allowing the use of Disney characters across franchises. If OpenAI stumbles, these downstream partnerships could unravel quickly.

The Diversification Safety Net

Yet even if OpenAI fails spectacularly, the damage to its major backers might be contained. Nvidia maintains partnerships with competitors like Anthropic, where it invested $10 billion as recently as November. Microsoft still has Office, Windows, and numerous other revenue streams.

"These aren't company-ending bets as far as I can tell," says Tomic. "If the stock takes a hit, it takes a hit, but Microsoft hasn't put so much money into OpenAI that its survival depends on it."

Sebastian Mallaby of the Council on Foreign Relations, who wrote a New York Times op-ed predicting OpenAI would run out of money within 18 months, sees a more targeted bubble: "There is a bubble, but not for AI, just for OpenAI. They don't have other products to fall back on."