The Wall Street Journal reported on the parallel this week, noting that the rise of Nvidia—whose share price has soared nearly 30,000% in the past decade—is reminiscent of the dotcom era when the hot new tech was not AI, but the internet. That era ended with a thud as the tech-heavy Nasdaq index fell over 50% in the year 2000 as Cisco and other once-hot stocks came down to earth. This doesn’t necessarily mean the current Nvidia-led boom will end in tears, too, though some are getting antsy.

“The insatiable flow of money into AI has raised eyebrows among investors uncertain the boom can continue without pause. Some $50 billion has been invested in Nvidia’s chips since the boom began, according to a Sequoia Capital estimate in March, but generative-AI startups have only brought in $3 billion in sales,” the Journal reported.

There are other reasons to believe Nvidia—along with other AI-tied stocks riding its coattails—is overvalued. As Fortune recently noted, the company’s current P/E multiple—a key valuation metric that looks at share price in light of earnings—is well over 100. This is way, way beyond the typical P/E for even a growth stock. For Nvidia to justify that multiple, it will have to post a long series of blow-it-out-of-the-water earnings reports of the sort no company has delivered before. A more likely outcome is its share price drops to reflect its actual earnings—and the broader market follows suit.

Robin Greenwood, a finance professor at Harvard Business School, says there are “certainly some echoes” in the current market to what took place in the dotcom era. But does this mean the outcome will be the same?

What we can learn from past bubbles

The nature of stock market bubbles is that they are much easier to see in hindsight. History shows that, in the midst of a boom, investors are more likely to be mesmerized by returns that keep climbing higher. That was the case in the internet-fueled dotcom bubble of 2000, and it was the case in 2008 when a proliferation of mortgage-backed securities drove stock indexes to new heights.

William Goetzmann, a finance professor at Yale School of Management, has studied the history of stock market bubbles—including the first one: the South Sea Bubble of 1720, when investors in London, Paris, and Amsterdam bid up shares in hopes of cashing in on a new era of overseas colonization. Eventually, the price paid for shares became untethered to economic reality, and a brutal crash ensued.

Goetzmann has an interesting twofold take on the South Sea Bubble and others that followed. The first is that, even though investors incurred heavy losses, share prices did not drop to pre-bubble levels. This, he says, reflects the fact that many of the companies selling shares were indeed helping to carry out a transformational change in society and the economy.

The same is true of the 2000 dotcom boom when Cisco and others were racing to bring broadband internet to millions of people and businesses. And while the 2008 boom was not based on a foundational new technology, it was fueled by novel forms of financialization that made it easier to divide and distribute investments.

“When you see something that looks like a bubble, as an economist, I always look at the innovation associated with it. How’s it going to change the way money is made or the way business is done?” Goetzmann says.

As for Nvidia’s soaring share price, he observes that “clearly there’s anticipation there will be another transformation”—this one based on AI changing daily life as much as the internet did 25 years ago. In this context, it’s no surprise investors are rushing into Nvidia even if there is likely some pain ahead.