Startupdaily
startupTelegram

Is the AI IPO Wave the Dot-Com Bubble in Disguise?

AI Markets·October 22, 2026

Is the AI IPO Wave the Dot-Com Bubble in Disguise?

Anthropic's proposed public listing, pegged at a valuation of around $2 trillion, has become a test case for how investors should judge the AI boom. In a guest column, Airrived co-founder and CEO Anurag Gurtu argues the math does not add up, and he draws a direct line to the late-1990s internet bubble. The column also frames a delay in the listing as evidence that the market is starting to ask harder questions.

Gurtu's central point is that fast revenue growth is not the same as a durable business. Frontier AI labs must lock in huge infrastructure commitments well ahead of demand, and those costs keep climbing whether or not customers follow. He also raises safety concerns, which could slow product releases or add expense, and customer concentration, where a small number of large buyers can account for a big share of income. Losing one or two of them would hit the top line hard.

Competition is the other pressure point. Open-weight models, which can be downloaded and run at lower cost, are closing the quality gap on many everyday tasks. If capable models become cheap and widely available, the premium that supports a sky-high valuation becomes harder to defend.

Gurtu's alternative is a different definition of value. He argues that lasting enterprise AI will come from proprietary data, reliable deployment inside real workflows, and measurable results for customers, rather than from owning the most expensive model. On that view, companies that sell outcomes may be better positioned than those that sell raw access to a model.

It is worth remembering that this is an opinion piece, and the $2 trillion figure is a proposal rather than a settled price. Still, the debate it raises is live across the sector. Whether the AI listing wave repeats the dot-com pattern will likely depend on whether revenue can keep pace with spending over the next few years.

Reporting based on an external source.