With a Web of Circular Deals, OpenAI and Nvidia Fuel a $1 Trillion Ai Market

Two weeks ago, Nvidia Corp. announced plans to invest up to $100 billion in OpenAI, the leading artificial intelligence startup, to help fund a massive data center expansion one so large it could theoretically power an entire city.

‍

In exchange, OpenAI committed to equipping those facilities with millions of Nvidia chips. The deal quickly drew criticism for what some analysts called its “circular” structure.

‍

Undeterred, OpenAI returned to the spotlight this week with another headline-grabbing agreement this time with Nvidia’s chief competitor, Advanced Micro Devices Inc. (AMD).

‍

On Monday, the ChatGPT creator unveiled a multibillion-dollar partnership to purchase tens of billions of dollars’ worth of AMD chips. As part of the deal, OpenAI will also become one of AMD’s largest shareholders, signaling its intent to diversify its hardware suppliers while keeping pace with the rapid growth of AI infrastructure.

‍

These massive investments underscore the speed and scale of spending in a technology that, despite its promise, remains largely untested as a consistent source of profit.

‍

Much of the money pouring into the sector can be traced back to just two players Nvidia and OpenAI whose expanding web of alliances and capital commitments is heightening fears of an AI-driven financial bubble. The implications stretch far beyond Silicon Valley, touching everything from global debt and equities to energy and real estate markets.

‍

Both companies were catalysts for the AI boom that began roughly three years ago, and they’ve kept the momentum alive through overlapping deals with cloud providers, AI developers, and startups. Now, some market watchers warn that this complex network of interdependence may be inflating the industry’s valuations and increasing the risk of a broader market correction. Collectively, OpenAI’s computing and chip agreements with Nvidia,

‍

AMD, and Oracle Corp. could exceed $1 trillion. Despite this, the AI pioneer remains deeply unprofitable and doesn’t expect to turn cash-flow positive until the latter part of the decade.

“If we look back a year from now and realize the AI bubble burst, this deal could be one of the early warning signs,” said Brian Colello, an analyst at Morningstar, referring to Nvidia’s massive investment in OpenAI. “If things go south, the problem may lie in these circular relationships.”

‍

‍

Nvidia’s aggressive investment strategy has long been an open secret in the tech world. The chip giant has funneled billions into dozens of AI startups most of which rely heavily on its GPUs to train and operate their systems. OpenAI, though less prolific in its investments, has also poured money into emerging firms building on top of its models. As the AI race moves into an even more capital-intensive phase, the size of these deals has ballooned, making the financial ecosystem increasingly intertwined.

‍

Just a day after announcing its $100 billion pact with Nvidia, OpenAI revealed a separate $300 billion partnership with Oracle to construct U.S.-based data centers. Oracle, in turn, has committed billions to purchasing Nvidia hardware for those facilities effectively channeling much of that money back to Nvidia itself.

‍

However, Oracle’s ability to profit from such vast spending has come under scrutiny. A recent report revealed that the company’s cloud division had thinner profit margins than analysts expected. Although Oracle generated about $900 million in revenue from renting Nvidia-powered servers last quarter, its gross profit was just 14 cents per dollar of sales, according to internal documents cited by The Information. The report sent Oracle shares lower, dragging on the broader market.

‍

Adding to the flurry of activity, sources said Tuesday that Nvidia plans to invest up to $2 billion in equity in Elon Musk’s xAI, as part of a $20 billion funding round that also includes roughly $12.5 billion in debt. The structure facilitated by a special purpose vehicle would see xAI purchasing Nvidia processors and leasing them out for five years.

‍

Similar financial loops are appearing elsewhere. CoreWeave Inc., one of the year’s standout tech IPOs, illustrates this dynamic perfectly. Nvidia acquired a 7% stake in CoreWeave ahead of its public debut and later agreed to spend $6.3 billion on CoreWeave’s cloud services. OpenAI, meanwhile, secured $350 million in equity from CoreWeave and expanded its own cloud contract with the company to as much as $22.4 billion further tightening the financial ties between OpenAI and Nvidia.

‍

Inside the industry, executives insist these arrangements are necessary to meet unprecedented demand for AI computing. AMD CEO Lisa Su described the OpenAI partnership as a “virtuous cycle,” while OpenAI President Greg Brockman said it will take a “full industry effort” to support ChatGPT and other AI products.

‍

In Washington, regulators appear content to let the market run. “It’s up to them,” said White House AI and Crypto Czar David Sacks, noting that U.S. companies must stay competitive with China. The Trump administration, meanwhile, remains indirectly linked through its stake in Intel Corp. and its intention to collect revenue from Nvidia and AMD’s chip sales to China.

‍

Still, some analysts see echoes of the late 1990s dot-com era. “Back then, circular deals were driven by startups buying each other’s ads and services to fake growth,” said Paulo Carvao, a senior fellow at Harvard Kennedy School. “Today’s AI firms have real products, but their spending still far outpaces revenue.”

‍

Nvidia insists that it doesn’t require any company it invests in to use its chips, while OpenAI declined to comment.

‍

Yet there’s no denying the staggering sums involved. OpenAI CEO Sam Altman has said the company could eventually invest “trillions” to build the physical backbone of next-generation AI an audacious goal for a firm that has yet to post a profit. To fund that vision, OpenAI plans to tap a mix of venture capital, debt, and strategic partnerships with firms eager to associate themselves with the AI leader.

‍

Altman addressed his company’s finances at a recent event: “One day, we’ll need to be very profitable, and we’re confident we’ll get there,” he said. “But right now, we’re focused on investment and growth.”

‍

Nvidia, on the other hand, has the balance sheet to sustain its expansion. With a $4.5 trillion market valuation and dominance in AI chips, the company has become the world’s most valuable firm. According to PitchBook, Nvidia participated in 52 AI-related venture deals in 2024 and nearly matched that number by September this year. CFO Colette Kress said the company aims to deploy capital into “the most strategic parts of the ecosystem.”

‍

CoreWeave CEO Michael Intrator, addressing concerns about circular financing, downplayed the issue: “When Microsoft buys infrastructure from us to serve its clients using Copilot or 365, I don’t care what people say about circular financing,” he said. “Real customers are using it that’s what matters.”

‍