The artificial intelligence boom increasingly resembles the dot-com era: the technology is genuinely changing the world, demand is growing, and companies are pouring astronomical sums into infrastructure. But there is still an enormous gap between AI's potential and the profitability of businesses built on it. Now, uncertainty surrounding Nvidia—a company that has become one of the industry's main financial and technological hubs—has added to the financial risks. Analysts warn that if investors and lenders begin losing confidence in large-scale AI projects en masse, a market correction could turn into a crisis with consequences far beyond the technology sector.

AI works—and that is precisely why a bubble could form around it

The current boom has an important distinction from many previous waves of technology speculation: artificial intelligence is not merely promising to transform the economy; it is already delivering practical benefits. Companies use it to write software, automate routine operations, and perform tasks that previously required substantial human effort.

But the reality of the technology does not guarantee financial success for every company seeking to profit from it. Veteran technology investor Paul Kedrosky summed up this contradiction in a conversation with Axios: no reasonable person doubts that AI works. And that is precisely what makes a bubble possible.

The problem lies in the scale of expectations. Technology companies are already committing trillions of dollars to AI development, increasingly relying on borrowed funds, cross-investments, and complex financing arrangements. Market participants expect demand for computing capacity to keep growing and investments in data centers and equipment to eventually pay off.

However, if the expected profits fail to materialize on schedule, these enormous investments may prove unjustified—even if the technology itself continues to advance.

Why uncertainty surrounding Nvidia is so dangerous

Nvidia plays a particularly important role in this situation. The company does more than sell AI accelerators: it has become a key participant in financing the infrastructure boom. Nvidia invests in its own customers and provides financial guarantees for major projects, helping them build data centers and purchase equipment.

Confidence in the company therefore affects not only demand for its chips but also the willingness of investors and lenders to finance the entire industry.

Against this backdrop, Jensen Huang's departure as Nvidia's CEO was a particularly consequential development. Chief Financial Officer Colette Kress took over as interim head of the company. The changes came at a time when the market was already facing financial and political uncertainty.

The departure of a leader who had long been one of the defining figures of the AI boom raised questions about whether Nvidia would continue supporting its partners' projects as actively as before. If the new management opts for a more cautious financial strategy, some initiatives could lose a crucial source of support.

The main risk here is not just a possible change in Nvidia's own policies. If lenders conclude that the company will no longer back additional large-scale projects, they may also decline to finance those projects themselves. RSM Chief Economist Joe Brusuelas believes the turmoil surrounding Huang could trigger a major bursting of the bubble. Columbia Business School professor Stijn Van Nieuwerburgh also warns that lenders' confidence in Nvidia's willingness to support new projects is critical to continued infrastructure construction.

Investment is growing faster than expected revenue

The AI market's problems, however, did not begin with the leadership changes at Nvidia. Financial tensions had been building for a long time.

According to estimates cited by Axios, AI service providers could spend about $7 trillion on infrastructure by 2030. Meanwhile, Gartner forecasts that AI-related revenue over the same period will not exceed $3.6 trillion.

This does not mean every investment will necessarily lose money: infrastructure can continue generating revenue after 2030, and revenue forecasts may change. But the gap between the scale of investment and expected returns shows just how much the market has wagered.

At the same time, valuations of the largest AI companies are being revised. According to Axios, the terms of Anthropic's planned IPO imply a business valuation well below earlier expectations, which had reached as high as $2 trillion. Media reports indicate that OpenAI also raised less money in a recent funding round, and at a lower valuation, than it had hoped.

Investor behavior is changing too. During the sharp decline in the U.S. stock market on November 5, retail traders who had previously been active dip buyers became net sellers for the first time in a long while. According to Vanda Research, this could signal a weakening of their usual willingness to support AI assets whenever prices fall.

Finally, financing itself is becoming more expensive. Rising borrowing rates and a large volume of new bond issuance are making lenders more cautious about projects that require enormous spending and promise returns only in the distant future.

Politics could also slow AI infrastructure construction

Political risks are adding to the financial ones. Opposition to new data centers is growing in the United States: candidates critical of AI infrastructure development are winning elections, while restrictions on where such facilities can be built enjoy significant voter support.

For the industry, this is not simply a matter of construction permits. Delays can increase project costs, disrupt plans to bring new computing capacity online, and undermine the financial assumptions underpinning investments.

There is also an international dimension. Following Huang's departure, calls from members of both U.S. political parties to tighten restrictions on advanced chip exports to China have intensified. If adopted, such measures could complicate Nvidia's business, although they could also limit Chinese competitors' access to the most powerful computing systems, thereby bolstering the position of American AI developers.

The industry thus faces pressure on several fronts at once: infrastructure construction could slow, lenders could become more cautious, and access to certain markets could be restricted.

How a financial crisis could spread across the industry

The most troubling scenario is a chain reaction. If AI companies' share prices begin to fall and borrowing becomes more expensive and harder to obtain, startups and cloud providers will find it more difficult to finance new data centers.

Less construction means lower demand for accelerators, servers, electricity, and other infrastructure components. This would hurt equipment suppliers and investors who have backed new projects. In turn, their deteriorating financial position could further restrict the industry's access to capital.

In this way, a stock market correction could become a much broader problem—especially if a substantial share of the infrastructure boom rests on borrowed money and expectations of continuously growing demand.

The technology itself, however, will not disappear. Financial difficulties will not change the fact that AI is already being used in real-world work and continues to advance. For example, Anthropic reported that its annualized revenue run rate had reached $81 billion by the end of October, up from $71 billion a month earlier. According to The Wall Street Journal, OpenAI increased its quarterly revenue by about 50% to $10 billion in the third quarter. However, the company's adjusted operating losses for the same period totaled $10.9 billion.

These figures clearly illustrate the contradiction in today's market: revenue is growing rapidly, but spending and financial expectations also remain enormous.

The dot-com lesson: technology can win while investors lose

The history of technological revolutions shows that the emergence of a truly groundbreaking technology is often accompanied by a speculative boom. In the late 1990s, investors poured money into internet companies, anticipating rapid growth in the new market. When the bubble burst, many companies disappeared and investors lost enormous sums. But the internet itself did not disappear: the infrastructure that had been built and the new business models became the foundation for the subsequent development of the digital economy.

MIT economics professor Ricardo Caballero believes excessive optimism is sometimes even necessary to attract enough funding to develop a new technology. He has no doubts about AI's transformative potential, but stresses that this does not rule out a financial bubble.

That is the central paradox of the current boom. Artificial intelligence may become one of the most important technologies of the 21st century, but that does not mean every company, investor, and lender betting on it today will come out ahead.

As Stijn Van Nieuwerburgh notes, AI can deliver enormous benefits to users even if the companies building the infrastructure for its development earn far less than their investors expected.

A potential market crash will not necessarily halt the artificial intelligence revolution. But it could certainly change its pace, its participants, and its sources of financing—and leave trillions of dollars in financial losses in its wake.