NEW YORK — Investors flocked to the artificial-intelligence initial public offerings this week, and the single most repeated compliment in the broadsheets was a phrase no one who has read a balance sheet can comfortably stand by: the new AI IPOs are, as one analyst put it, characterized by “big dreams and tiny revenue.” That is, the companies are being valued in the hundreds of billions on the strength of projections, which is a valuation strategy that has not, historically, required the company to have much of a revenue line.

The contrast, as the Financial Times’ Lex column put it bluntly, is “striking” — a word that in market commentary is doing the work of a spreadsheet. The most-watched offering of the year, the long-awaited debut of a frontier model maker with roughly $2 billion a month in revenue and a projected $14 billion loss for 2026 at a valuation that crosses the $1 trillion mark, is now the reference point every other offering is measured against, the way the weather is the reference point every other conversation.

"The nature of the AI boom and the size of the valuations make the contrast between ambition and income more striking," the Financial Times wrote, which is the closest the market has come to a disclosure requirement all year.

The numbers, when you lay them out without a valuation next to them, are a little hard to hold in the same hand as the price. The lead offering is projected to lose roughly $14 billion in 2026, is not expected to be cash-flow positive for years, and is priced, according to the people buying it, at a level that would require the revenue line to catch up to the valuation line at a speed the company itself has not committed to. The investors, for their part, are describing this not as a risk but as a “thesis.”

THE AI IPO, PER THE BROCHURES

  • Revenue: enough to be real, not enough to be the price
  • Projected 2026 loss: a nine-digit number with a friend
  • Valuation: a ten-digit number with a friend
  • Cash-flow positive: "eventually," per the roadshow
  • The compliment Wall Street gave it: "big dreams"
  • The other compliment: "tiny revenue"

The skeptics, a small crowd near the coffee machines, point out that a valuation is, by definition, a claim about the future, and that the future has a way of not being a subscription business. The enthusiasts, a larger crowd holding more of the portfolio, point out that the skeptics were also skeptical about the previous four of these offerings, and that the skeptics’ funds have, at this point, underperformed a parking meter.

At press time, the offering’s pricing round was still open. The underwriters, asked whether the revenue line would need to do anything before the valuation line made sense, replied that the revenue line was “doing more than most people think,” which, in the lexicon of a roadshow, is a sentence with the exact same information content as “we think it’s going to be fine.”