The AI boom has not necessarily ended, but parallels with the dotcom market in summer 2000 merit caution, according to Dario Perkins, macro economist at GlobalData.TSLombard.

“For the most part, we have maintained a constructive tone. In fact, there are good reasons to think the AI boom still has further to run,” Perkins wrote in an analysis research note, despite recent pressure on AI stocks.

Access deeper industry intelligence

Experience unmatched clarity with a single platform that combines unique data, AI, and human expertise.

Find out more

Major hyperscalers have yet to reduce their capital expenditure plans, while demand for computing capacity remains solid, as indicated by GPU rental prices. The Fed has also only recently started a new tightening cycle.

Perkins added that market history points to another potential source of support: “there is always a market correction – followed by a prompt rebound – ahead of the ultimate crash.”

However, he said the opening phase of the Dotcom crash offers a warning against relying too heavily on these arguments. “The arguments we are making today about why the boom will keep going are eerily similar to the arguments people were making as the Dotcom bubble burst,” Perkins wrote.

During the summer of 2000, the Nasdaq recovered 35% from its lows by mid-June, leaving it only 15% below its March peak. Investors continued to buy market declines, drawing confidence from the gains generated by the same strategy during the late 1990s.

At the time, the prevailing view was that technology groups were “relatively immune to higher interest rates”, given lower debt levels and rapidly increasing earnings expectations. Investors also moved towards businesses with “real earnings”, including Cisco, Oracle, Sun, Intel, and Microsoft.

That recovery proved incomplete. By mid-October, the Nasdaq had fallen below its previous lows, and it ended 2000 at about half its March peak after dropping more than 40% in three months.

Perkins does not conclude that the current AI cycle has peaked. But the Dotcom experience shows that market confidence, strong demand, and visible technological change do not remove the risk that valuations could later come under pressure.