Stocks of companies related to artificial intelligence have already experienced a decline of 20–50%, but the current market growth remains an order of magnitude more modest than what preceded the dot-com bubble crash in 2000. A comparative analysis of the two cycles shows: the AI rally lags behind the dot-coms on all key metrics, which means a classic "collapse" is still far off.
The analysis is based on two starting points. The dot-com boom is counted from the release of the Netscape browser in December 1994, and the AI boom from the launch of ChatGPT in November 2022. A comparison across three parameters provides a clear picture.
Smaller scale, shorter duration
The dot-coms grew for 5.3 years, whereas the AI rally has been going for only 3.9 years. If the past cycle were simply repeated, there would still be about 1.4 years of growth ahead.
The index dynamics are even more telling. During the dot-com era, the Nasdaq 100 grew by roughly 1000%. After the launch of ChatGPT, it has gained only 137% — just 13% of the previous result. The peak in March 2000 was followed by an 83% decline in the index over two and a half years — precisely the outcome investors fear as they wonder whether the AI bubble has burst.
Nvidia is far cheaper than Cisco at its peak
Cisco and Nvidia have become the symbols of the two eras. 3.9 years after the start of the cycle, Cisco had grown by roughly 700%, and at its peak its growth reached 4200%. Nvidia at the same point in time has gained about 1000% — more, but with a far more modest valuation.
The key difference lies in the forward P/E multiple. At its peak in March 2000, Cisco was valued at 131 times expected annual earnings, whereas Nvidia currently trades at 16.9 with a market capitalization of $4.9 trillion. Simply put, P/E shows how many years a company must earn its expected profit to recoup its price: the higher the multiple, the more the stock depends on hopes rather than actual income. Cisco's figure was almost 8 times higher than Nvidia's today. Cisco was priced "for a perfect scenario," while Nvidia is priced like an ordinary mature industrial company.
The dot-com bubble burst when stocks were very expensive relative to earnings and the rally had lasted too long. Right now, neither is the case. The current drawdowns are not a reason to sell everything, but an ordinary correction within a rising market.
Cryptalist analysis: The comparison of the two cycles points to a structural difference in the quality of the current rally: Nvidia's moderate multiples and the index's modest gain relative to the 1990s reduce the likelihood of a systemic collapse along the 2000 scenario. Nevertheless, the high concentration of capital in a narrow group of AI leaders makes the market vulnerable to corrections at any disappointment over the pace of technology monetization. Investors should assess risks through the lens of valuations, not emotions around the word "bubble."