On 10 March 2000, the Nasdaq peaked at 5048.6 points – a moment that no one recognised as such at the time, but which, in retrospect, marked the beginning of a freefall that would wipe out 77 percent of the market’s value. Now, 25 years later, inevitable parallels arise between the dotcom hype and the current AI revolution. But are these comparisons justified?
If a stock market is predictable at all, the peak is far harder to time than the bottom. Fear clusters into a panicked sell-off, but euphoria can persist for much longer – and usually does. The smartest analysts on Wall Street, who were ultimately proven right, predicted the peak far too early. Admittedly, central bankers are usually even earlier, although it remains debatable whether Greenspan’s December 1996 “irrational exuberance” speech should be seen as a warning or an encouragement.
Only when these lone voices crying in the wilderness are immediately vindicated by a crash does such a market collapse grant them a lifetime of lucrative speaking engagements. The market is merciless when it comes to timing. In December 1999, partners at a renowned American investment bank went short on tech stocks, only to be squeezed out like amateurs two months later.
The unpredictable peak
A stock market bubble follows a predictable pattern: the peak is virtually impossible to time, whereas a bottom is often easier to identify. In 1999, even Wall Street’s sharpest analysts issued their warnings months too early. The famous “irrational exuberance” speech by Fed Chairman Greenspan had, moreover, been delivered as early as December 1996 – three and a half years before the eventual crash.
What made the dotcom bubble so treacherous was the collective euphoria in its final phase. Amid this frenzy, Lehman Brothers quadrupled Softbank’s price target to 400,000 yen simply because its share price had doubled in one week. World Online went public with valuations a hundred times higher than recent private transactions. The Financial Times in March 2000 raved about the “phenomenal Nasdaq” reaching “dizzying heights” – even weeks after the crash had already begun.
AI and tech in 2025: echoes from the past?
This history sounds ominous in light of recent developments. The S&P500 has recently reached new records, but global markets are now in turmoil. US consumer confidence is declining, manufacturing orders are plummeting, and the S&P500 has already given up its post-election gains.
Most concerning are the so-called Magnificent Seven tech giants – Apple, Microsoft, Alphabet, Amazon, Nvidia, Meta, and Tesla – which are showing worrying signs. These stocks, responsible for the lion’s share of market gains in recent years, have now corrected by 12 percent since their December peak. And this despite strong quarterly earnings, such as double-digit growth at Alphabet. Investors are beginning to ask: when will the billions invested in AI finally translate into tangible growth?
Fundamental differences
Despite some parallels, there are fundamental differences between the dotcom companies of 2000 and today’s Magnificent Seven tech giants. The Mag7 (Apple, Microsoft, Amazon, etc.) are not fragile startups but globally profitable enterprises with proven business models and strong cash flows. They are valued based on real financial performance, not speculative metrics like “eyeballs” or “clicks-per-minute.”
Unlike in 2000, when the internet was still in its infancy, AI is being built upon an already existing digital infrastructure. Whereas dotcom companies burned through venture capital just to establish basic functionality, the Mag7 invest in AI from their own cash flows. Moreover, dotcom firms rarely had clear revenue models, whereas today’s tech giants operate lucrative subscription models that, combined with their quasi-monopoly positions, are highly profitable.
The rise of Chinese AI models such as DeepSeek is accelerating AI application growth, which is often integrated into existing profitable products. The Mag7 are also far more dominant than the thousands of fragmented dotcom companies, most of which no longer exist.
The regulatory landscape has also reversed: after the dotcom crash, scandals such as Worldcom and Enron led to stricter regulations, whereas there is now pressure to deregulate. Monetary policy differs as well: at the end of 1999, central banks were raising interest rates, whereas now they are lowering them. However, the most important distinction is valuation. Extreme valuations are characteristic of bubbles, yet the current valuation of the Mag7 is far below that of previous bubbles such as the dotcom, Japanese, or Nifty Fifty bubbles.
Conclusion
History rarely repeats itself exactly. While we must always remain vigilant for signs of a potential bubble, the fundamental differences suggest we are not necessarily heading for a dotcom-like implosion. The valuations of leading AI companies are supported by substantial profits and cash flows, not vague promises. Ironically, the internet ultimately proved even more valuable than the most optimistic dotcom speculators could have imagined in 2000 – but not for the companies that dominated headlines at the time.
Whether AI will have a similar long-term impact remains to be seen. But one lesson from 2000 still holds: only when “everyone” is euphoric about an investment trend is it perhaps the most reliable signal to seek the exit. Do not do this by going liquid, but by diversifying into strongly undervalued markets. Given the strong market concentration of recent years, there are plenty of those available.
Han Dieperink is Chief Investment Officer at Auréus Vermogensbeheer. Earlier in his career, he was Chief Investment Officer at Rabobank and Schretlen & Co.