Goldman Sachs says AI boom resembles 1997 tech cycle, not dotcom peak, with room to run

Goldman Sachs says the artificial intelligence investment cycle still has room to run, comparing today’s surge in AI spending and valuations to the early phase of the late-1990s tech boom rather than its speculative peak.

In a new report, the bank said that “on many metrics, the current AI-related boom has more in common with the tech boom in 1997/1998 than in 1999 or 2000.” That period marked the buildout stage of the internet era, when productivity and infrastructure gains were beginning to emerge but before market excesses took hold.

Goldman cautioned that strong returns on capital are not guaranteed, but said current trends suggest “plenty of room for the AI investment boom to run” provided no external shocks or funding constraints disrupt momentum.

The analysis implies that AI’s expansion phase could continue as companies invest heavily in data centres, semiconductors, and model training infrastructure. The bank’s framing reinforces the view that AI remains in its buildout stage — with parallels to the pre-dotcom frenzy years when long-term productivity gains were only starting to materialise.

—

Goldman’s comparison to the pre-dotcom years reinforces optimism in AI-linked equities and infrastructure plays. The bank’s view could bolster sentiment across chipmakers, cloud providers, and data-centre developers seen as core to the current investment wave.