On September 22, 2025, Nvidia, the world’s most valuable semiconductor company, made headlines with its announcement of a colossal $100 billion investment in OpenAI, the creator of ChatGPT. The deal, which aims to build a 10-gigawatt (GW) AI data center—equivalent to the power output of ten nuclear plants—has ignited a fierce debate across Wall Street, Silicon Valley, and beyond. Some see it as a bold leap into the next era of artificial intelligence, while others raise alarms about echoes of past financial bubbles.
The core of Nvidia’s plan is to enable OpenAI to train and deploy its advanced models using Nvidia’s state-of-the-art AI chips. The scale is staggering: the data center is expected to require 4 to 5 million Nvidia GPUs, matching the company’s entire projected chip shipments for 2025 and doubling last year’s output. According to Bloomberg, this partnership could fundamentally reshape the AI landscape, but it’s not without controversy.
Critics have been quick to draw parallels to the infamous dotcom bubble of the early 2000s. Back then, companies like AOL propped up their revenues through “round-trip” transactions—essentially paying customers to buy their own products, artificially inflating financial results. Regulators eventually cracked down, and AOL paid a $300 million penalty to settle fraud claims. Today, skeptics argue that Nvidia’s investment in OpenAI could be a modern-day version of such circular financing. As Nvidia pours money into OpenAI, OpenAI is expected to spend billions on data centers stocked with Nvidia chips, creating what some analysts call a “vendor financing” loop.
Stacy Rasgon, an analyst at Bernstein Research, didn’t mince words: “This partnership will clearly stoke ‘circularity’ concerns.” Jay Goldberg of Seaport Global Securities echoed this sentiment, describing the transaction as symbolic of “bubble-like behavior.” Vivek Arya of Bank of America added, “The optics of such a large investment in a customer will raise questions until Nvidia clarifies the appropriate accounting treatment.” Even OpenAI’s CEO, Sam Altman, admitted last month, “Are we in a phase where investors as a whole are overexcited about AI? My opinion is yes.”
It’s not just Nvidia and OpenAI raising eyebrows. The AI investment boom has triggered a cascade of similar deals across the tech sector. Earlier this month, Oracle, led by billionaire Larry Ellison, struck a $300 billion deal with OpenAI that also involves hefty purchases of Nvidia chips. Amazon, not to be left behind, has invested billions in Anthropic, an OpenAI rival that relies heavily on Amazon’s data centers. Nvidia itself has backed a string of AI startups, including a $500 million commitment to British firms Wayve and Nscale, both dependent on Nvidia hardware. ASML, the Dutch microchip manufacturing giant, recently pledged €1.3 billion to French AI startup Mistral.
This spending spree is having profound effects on the broader economy. According to Gartner, global data center expenditures will soar by 42% this year, reaching $474.8 billion. Deutsche Bank analysts went so far as to claim that the U.S. “would be close to, or in, recession this year” without the AI-driven data center boom. “AI machines—in quite a literal sense—appear to be saving the US economy right now,” they observed. The scale of investment is almost hard to fathom, but the big question remains: will the returns justify the outlay?
Some experts are skeptical. OpenAI’s annual revenues currently stand at $12 billion, while S&P forecasts global generative AI revenues to hit $30 billion in 2025 and $85 billion by 2029. Even under the most optimistic scenarios, Bain & Company estimates that AI-related revenues would still be $800 billion short of what’s needed to fund all the planned infrastructure spending by 2030. Aswath Damodaran, a finance professor at New York University, summed up the concern: “Way too much is being spent on AI infrastructure, given that the market for AI products and services is still at a hope and pray stage.”
Doubts about the immediate payoff from AI investments have been reinforced by recent academic studies. In August, researchers at the Massachusetts Institute of Technology reported that 95% of organizations deploying AI saw “zero return.” The news triggered a brief panic that wiped $1 trillion off U.S. tech stocks. Just this week, a Stanford and BetterUp Labs study found that AI might even be harming productivity, as office workers churn out polished but ultimately vacuous reports—what the researchers dubbed “workslop”—forcing colleagues to spend hours cleaning up the mess.
Yet, not everyone is convinced a bubble is about to burst. Dan Ives, a well-known analyst at Wedbush Securities, sees the Nvidia-OpenAI deal as a pivotal moment. “The next three to six months will be when the second-, third-, and fourth-order effects of the AI revolution kick into full gear,” he predicted. Ives likened today’s environment to 1996, the early days of the internet’s explosive growth, rather than 1999, when the dotcom bubble reached its unsustainable peak. “While there are worries about an ‘AI bubble’ and stretched valuations, we continue to view this as a 1996 moment for the tech world and NOT a 1999 moment,” Ives said. In other words, he believes the AI sector is just getting started, with plenty of room to run before any crash looms on the horizon.
Importantly, the AI bubble debate isn’t confined to the U.S. The news has rippled through global markets, particularly in South Korea, where the semiconductor sector has led a recent rally. Foreign investors are watching closely, aware that the outcome of this debate could sway their strategies and impact the fortunes of entire economies. As the value and risks of AI investments come under scrutiny, markets worldwide are bracing for possible volatility.
Meanwhile, the debate over circular financing and speculative excesses continues to rage. Dom Rizzo, a portfolio manager at T Rowe Price, offered a nuanced view: “AI has the potential to be the biggest productivity enhancer for the global economy since electricity,” he said, but cautioned that “productivity cycles historically are accompanied by speculative bubbles.” Rizzo pointed to the proliferation of circular deals and the explosion of debt used to finance data centers as signals that “we are entering the next phase of the bubble.”
As history has shown—from the dotcom bust to the collapse of Enron and Ireland’s property bubble—market manias often end abruptly. Whether the current AI boom will follow that pattern remains to be seen. For now, investors, technologists, and policymakers are left to weigh the risks and rewards, hoping to distinguish genuine innovation from financial sleight of hand.
In the months ahead, all eyes will remain on Nvidia, OpenAI, and their peers, as the world waits to see whether this AI revolution will deliver on its promise—or prove to be another cautionary tale in the annals of financial history.