AI stocks have had a wild ride over the last month. First, they fell over concerns about a tightening credit environment, and then, following the midterm elections, fears of a crackdown on data centers sent the Invesco QQQ Trust (QQQ +2.31%) down 6.5% on Nov. 5.

However, since then, the ETF that tracks the tech-heavy Nasdaq-100 has clawed back nearly all of those losses, a remarkable comeback in just four sessions.

In fact, in the history of the ETF, which was launched in 1999, that type of event has only happened a handful of times, and none of those have taken place in the last 25 years. There have been four times when the Invesco QQQ Trust has fallen more than 5%, followed by three sessions that included a gain of at least 1%, two of which were more than 2% gains, and no sessions that included a decline of 1% or more.

The four dates when this pattern occurred before were:

• Feb. 3, 2000
• May 16, 2000
• Jan. 11, 2001
• April 20, 2001

You’ll notice that all of those dates came during the dot-com bubble or subsequent bust. In fact, a drawdown of at least 59% followed every one of those dates.

The Nasdaq-100 peaked during the dot-com era with a closing high of 4,704 on March 27, 2000, and didn’t hit bottom until Oct. 7, 2002, at a closing low of 807.42, a drawdown of 83% from the peak.

Is this the beginning of the AI bubble bursting?

Nearly since the beginning of the AI boom, investors have drawn comparisons to the dot-com era, which saw the Nasdaq-100 jump more than 700% from August 1995 to its peak.

There are a number of similarities with that period, including the early stages of a breakthrough technology, a surge in tech stocks, and a ramp in capital expenditures to build infrastructure to support the new technology.

However, there are also important differences that may explain why the recent volatility in the QQQ is different from what occurred during the dot-com bust.

First, the QQQ’s valuation is much lower than it was during the height of the dot-com boom. Currently, the ETF trades at a price-to-earnings ratio of just 28, making it only slightly more expensive than the S&P 500. By contrast, the QQQ’s PE ratio was approximately 82 when it peaked during the dot-com bubble, or three times more expensive than it is today. The Nasdaq has also not even tripled from its bottom in 2022, meaning its gains are much smaller than they were in the dot-com era when the Nasdaq-100 doubled in a single year, 1999.

Additionally, the AI boom is largely being driven by established, highly profitable companies like Nvidia, Alphabet, and Microsoft that are justifiably worth trillions of dollars. On the other hand, the dot-com boom was characterized by hundreds of unprofitable start-ups flooding the market, which helped drive the bust once it became clear that their business models weren’t viable. In the current era, there have been few IPOs. However, there is some concern about the business models of OpenAI and Anthropic, the AI labs at the center of the boom, though Anthropic has now reported an adjusted operating profit. OpenAI, on the other hand, is still deeply unprofitable.

Will history repeat itself?

Of the four dates above, only one came before the bubble burst, and that plunge was caused by strong economic data that was expected to trigger an interest rate hike from the Federal Reserve.

That’s not unlike the current situation where the Fed is expected to raise interest rates due to stubborn inflation, and investors are worried that that could squeeze the debt-fueled infrastructure build-out at the center of the boom. AI stocks jumped on Wednesday, in part because CoreWeave was able to tame concerns about a debt bubble in AI infrastructure and its own solvency.

Overall, while there are similarities between now and the dot-com era, I don’t think we’re on the verge of a bubble bursting the way the market was then.

Demand for AI infrastructure still greatly exceeds supply, as CoreWeave just reminded investors, and valuations are still reasonable or even cheap, reined in in part by the psychic scar of the dot-com bust.

Overall, investors can learn a lot from studying past market cycles and should be aware of the risks facing the AI boom, including debt-fueled spending and the cyclicality of chip demand. However, the facts on the ground don’t indicate that any bubble is on the verge of bursting, despite the volatility in the QQQ.

AI demand remains strong, and even CoreWeave, the most vulnerable AI infrastructure stock, just reassured investors that it can fund its ongoing expansion. The AI boom still looks healthy.