On March 10, 2000, the Nasdaq Composite index peaked at 5,048.62, more than double its value from just a year earlier. Internet startups with “dot-com” slapped on the end of their names and little more than ideas to sell to investors had soared to astronomical valuations over the previous five years.
But all parties come to an end.
Within weeks of that high, the index began its fall, with a mix of factors spooking investors. Eventually, the Nasdaq shed more than 75% of its value, wiping out $5 trillion in market capitalization. The age of irrational exuberance was over.
A quarter-century later, the rise of AI has market veterans debating whether history repeats, or merely runs the same bubble algorithm with new inputs. Some market watchers see structural differences preventing an immediate collapse, while tech critics warn we’re debugging the same old bubble logic. Beyond balance sheets lies a deeper concern: We keep forgetting that the social costs of tech revolutions outlast their market corrections.
There are similarities between the market today and 25 years ago as some investors grow nervous, according to Kairong Xiao, a finance professor at Columbia Business School. But if the moment we are in ends up bursting, the market structure means things will look different than 2000. “The dot-com bubble was characterized by a lot of IPOs, with companies trying to get into the biggest asset class in the Nasdaq” Xiao said. “Today, due to regulation and many other structural changes, new investments nowadays have been through the private equity market.”
That structural shift from public to private funding represents one of the most significant changes. While Netscape’s 1995 IPO sparked a gold rush of public offerings from internet startups, today’s most prominent AI developers — OpenAI and Anthropic — remain private despite valuations in the tens of billions. They can access enormous capital without facing public market scrutiny or quarterly earnings pressure.
The companies leading today’s tech wave also differ markedly from their dot-com predecessors. In 2000, many market darlings were freshly minted startups with minimal revenue. Today, AI innovation is largely funded by cash-rich tech giants including Microsoft (MSFT), Google (GOOGL), and Nvidia (NVDA), companies with diverse revenue streams and substantial profits. This concentration of resources among established players may provide a cushion against market volatility that didn’t exist during the dot-com era.
“The big companies can survive if there’s a crash in the market because they have other ways to raise financing, whereas the small guys get eliminated,” Xiao said.
Yet not all market observers are convinced we’ll avoid a reckoning entirely. Gene Munster, managing partner at Deepwater Asset Management, said today’s AI boom has runway left — but also an expiration date.
“I still think we got another two good years left in the market,” Munster said. “Unfortunately, it’s gonna end in the same way that it did 25 years ago.”
For Munster, the key indicator isn’t current valuations but the psychological state of the market, which he sees as “euphoric” right now. He defines euphoria — that intoxicating moment when rational investing gives way to the fear-of-missing-out — as when the forward price-to-earnings multiple exceeds 50 times. Today’s Nasdaq trades around 25 times earnings. But Munster said that euphoric threshold could be reached by 2027, potentially triggering what he described as a “spectacular burst.”
While Munster said he believes the substance of AI will eventually exceed the hype, that likely won’t happen in time to avoid a market correction.
“By the time the numbers start to improve, the bubble will have burst,” Munster said. “I lived through the dot-com bubble. I don’t want to see people’s lives get blown up, but I also think that AI is going to be more impactful than the internet.”
Yet for all the structural differences between eras, some experts warn that we risk repeating the same fundamental mistakes. Mar Hicks, a history of technology professor at the University of Virginia and author of “Programmed Inequality,” said our retrospective view of previous tech bubbles often minimizes their negative impacts.
Those harms, Hicks said, include substantial erosion of privacy norms and labor market transformations that began during the first internet boom. Early e-commerce companies like Kozmo.com used venture capital to offer services at unsustainably low prices, conditioning consumers to expect convenience without cost — a precursor to today’s gig economy.
“That led to getting people used to this idea that you can have anything for free, essentially,” Hicks said. “It made it so that people either didn’t think about the labor and paying the labor of the people who are doing this work, or it completely changed the structure of the labor market.”
Hicks suggested that while Silicon Valley’s memory may be short, its executives aren’t entirely forgetting history — they’re strategically misremembering it. Today’s AI leaders appear to have studied the dot-com collapse primarily as a public relations case study rather than a substantive business lesson.
“The way in which they’re being mindful is they now know what they have to say in order to maybe get people to not get upset about these new revolutionary technologies,” Hicks said. “They’re giving lip service to it because they don’t ever address any of the real problems.”
The post The dot-com bubble burst 25 years ago. What does that history say about stocks today? appeared first on Quartz.
