When Jensen Huang, the chief executive of the chipmaker Nvidia, met with Donald Trump in the White House last week, he had reason to be cheerful. Most of Nvidia’s chips, which are widely used to train generative artificial-intelligence models, are manufactured in Asia. Earlier this year, it pledged to increase production in the United States, and on Wednesday Trump announced that chip companies that promise to build products in the United States would be exempt from some hefty new tariffs on semiconductors that his Administration is preparing to impose. The next day, Nvidia’s stock hit a new all-time high, and its market capitalization reached $4.4 trillion, making it the world’s most valuable company, ahead of Microsoft, which is also heavily involved in A.I.
Welcome to the A.I. boom, or should I say the A.I. bubble? It has been more than a quarter of a century since the bursting of the great dot-com bubble, during which hundreds of unprofitable internet startups issued stock on the Nasdaq, and the share prices of many tech companies rose into the stratosphere. In March and April of 2000, tech stocks plummeted; subsequently many, but by no means all, of the internet startups went out of business. There has been some discussion on Wall Street in the past few months about whether the current surge in tech is following a similar trajectory. In a research paper entitled “25 Years On; Lessons from the Bursting of the Technology Bubble,” which was published in March, a team of investment analysts from Goldman Sachs argued that it wasn’t: “While enthusiasm for technology stocks has risen sharply in recent years, this has not represented a bubble because the price appreciation has been justified by strong profit fundamentals.” The analysts pointed to the earnings power of the so-called Magnificent Seven companies: Alphabet, Amazon, Apple, Meta, Microsoft, Nvidia, and Tesla. Between the first quarter of 2022 and the first quarter of this year, Nvidia’s revenues quintupled, and its after-tax profits rose more than tenfold.
The Goldman paper also provided a salutary history lesson. Between 1995 and 2000, it pointed out, the tech-heavy Nasdaq index rose fivefold, and at the peak of the market a widely used valuation measure for the stocks that trade on it—the price-to-earnings ratio, or “P/E”—topped a hundred and fifty, a level not seen before or since then. By comparison, the five-year period from March, 2020, to March, 2025, had been relatively tame. It’s true, the Nasdaq had roughly doubled, and the P/E ratio had gone up considerably; but it hadn’t got anywhere near three figures.
Having written extensively on the dot-com boom and bust, I found some of Goldman’s analysis persuasive. Many people have either forgotten, or are too young to remember, the extremes reached during the dot-com years. In the logic of speculative hysterias—from the seventeenth-century “tulipmania” in Holland to the rise of Pets.com—greed, FOMO, and the greater-fool theory of investing eventually combine to banish caution, common sense, and financial gravity. Back in March, there was plenty of FOMO and trend-following on Wall Street, but it hadn’t reached the levels of the late nineties. Five months on, however, echoes of the dot-com era are getting louder.
Consider Palantir Technologies, whose A.I. software is used by the Pentagon, the C.I.A., and ICE, not to mention by many commercial companies. A couple of days before Huang visited the White House, Palantir released a positive earnings report. By the end of the week, according to the Yahoo Finance database, the market was valuing the company at more than six hundred times its earnings from the past twelve months, and at about a hundred and thirty times its sales in that same time span. Even during the late nineties, figures like these would have raised eyebrows.
Eye-popping I.P.O.s, another feature of the dot-com era, are also making a comeback. At the end of July, Figma, a firm that makes software used by internet developers, and which has added A.I. features to its suite of products, issued stock on the New York Stock Exchange at thirty-three dollars a share. When trading started, the price jumped to eighty-five. It closed the day at $115.50—a two-hundred-and-fifty-per-cent gain on the offering price. Watching this market action, I was reminded of August 9, 1995, when Netscape, which made the Netscape Navigator web browser, went public. Its stock was priced at twenty-eight dollars, rose to seventy-five, and closed at $58.25. In percentage terms, this leap was smaller than the first-day rise in Figma’s stock, but it’s often described as the beginning of the dot-com bubble.
It should be noted that, since Figma’s I.P.O., its stock has fallen back to below eighty dollars. This could be interpreted as a sign of sanity prevailing, but, given that the shares are still trading at more than double the offering price, other privately owned A.I. companies will be encouraged to enter the stock market. Renaissance Capital, a research firm that specializes in I.P.O.s, lists eight prominent candidates: OpenAI, Anthropic, Cohere, Databricks, SymphonyAI, Waymo, Scale AI, and Perplexity. Almost all of these companies are unicorns: they have been valued at more than a billion dollars in fund-raising deals with venture capitalists and other early investors. But, across the country, according to the research firm Tracxn, there are about seven thousand smaller and lesser-known A.I. companies, more than a thousand of which have already received Series A funding from external backers to finance their operations.
The ready availability of early-stage funding means that a necessary condition for a dot-com-style bubble is in place. So are three more: excitement among investors about a pathbreaking technology—generative A.I. clearly has the potential to impact great swaths of the economy; a Wall Street production line staffed by investment bankers eager to earn fees for organizing I.P.O.s; and accommodative policy. Last month, the Trump Administration announced an “AI Action Plan,” which aims to remove barriers to the deployment of the new technology and to deter individual states from introducing “burdensome” regulatory A.I. laws. The Federal Reserve, meanwhile, appears to be preparing to cut interest rates next month, which could give another boost to the markets.
There are, however, some important differences between now and the nineties, one of which is that the online economy is no longer a vast open plain on which enterprising individuals can propose to build castles to the sky. It is a redoubt of monopoly capitalism, in which Big Tech dominates the horizon. During the dot-com era, or its early stages, anyway, small startups could reasonably hope to exploit first-mover advantage, gain early traction, and create durable business franchises. In the A.I. economy, it seems possible that many of the rewards will go to top firms that can afford to build and maintain large A.I. models and can use their market power and financial might to ward off, or buy out, potential competitors. A vigorous antitrust policy could perhaps prevent this from happening, but, as the Wall Street Journal reported last week, the Administration’s pledge to pursue such a policy is now under threat from lobbyists and power brokers with close ties to the President. If investors decide that monopolies are the future of the A.I.-driven economy, the outcome in the stock market could well mean further gains for existing industry giants rather than a broad-based bubble.
All of this is uncertain, of course. The A.I. boom is still in the stage of building out infrastructure—training large language models, building data centers, and so on. A.I. applications are just beginning to diffuse throughout the economy, and nobody knows for sure just how transformative, and profitable, the technology will be. In this environment, many investors are following the time-honored gold-rush strategy of buying the shovel-makers and big mine owners. But history teaches us that even this strategy is far from risk-free. In an interesting analysis that was posted on the financial-news platform Seeking Alpha, an analyst identified as KCI Research compared Nvidia to Cisco Systems, one of the firms whose stock went parabolic in 1998-99. Just as Nvidia’s G.P.U.s (graphics-processing units) are now widely regarded as must-have components of A.I. infrastructure, Cisco’s routers and other network equipment were viewed as essential components of the internet build-out; for a time, demand for them seemed virtually unlimited. Like Nvidia, Cisco was an innovative and highly profitable company. But, in April of 2000, its stock dropped by almost forty per cent, and a year later it had fallen by about eighty per cent. A quarter of a century on, it still hasn’t recovered the high it hit in early 2000, although, lately, it has come close.
The Nvidia-Cisco comparison was a useful reminder of a dictum from the pioneering stock analyst Benjamin Graham, who was a mentor to Warren Buffett: in the short run, the stock market is a voting machine, but in the long run it is a weighing machine that weighs the cash flows that companies generate. Ironically, the Nvidia-Cisco analogy also inadvertently demonstrated how long the short run can last for, and how dangerous it can be to predict its end date. The analysis was posted in February of last year. Since then, Nvidia’s stock price has risen by another hundred and fifty per cent. ♦