For most of the AI boom, the two leading labs have been private companies spending extraordinary sums of other people's money. That is about to change — starting with Anthropic, which is preparing to file publicly for an initial public offering as soon as the end of this month. OpenAI is not far behind on paper: its confidential filing is already in front of regulators, though its own CFO has pointed to a 2027 listing rather than an imminent one. Between them, the two defining companies of the AI era are now both on the path to the public markets — at a scale that has no real precedent.

How big we're talking

Anthropic's ambitions are the eye-catching ones. The company is reportedly running the numbers on an offering that would match or exceed SpaceX's record-setting IPO — which would make it, on some projections, the largest public offering in history. That is against a valuation of roughly $965 billion set in its Series H round earlier this year. OpenAI, whose listing looks further out, would still command a debut that various reports place anywhere from the mid-hundreds of billions to north of a trillion dollars.

Numbers that size stop being ordinary corporate finance and start being macroeconomic events. An IPO of that magnitude reprices an entire sector, sets a public benchmark that every other AI valuation gets measured against, and — for the first time — forces the most closely watched companies in technology to open their books.

The opening of the books is the real story

That last part is what makes this more than a headline. A private company can brief reporters on a rosy forecast; a public one has to put it in a prospectus and stand behind it. When we looked at Anthropic's IPO case a couple of weeks ago, the hard question was whether the pitch — a valuation resting heavily on projected revenue years out — would survive contact with an actual filing. We are about to find out.

The two go into this from very different financial places, and the prospectuses will make that plain. Anthropic has just turned its first-ever quarterly operating profit — a milestone most of the field is nowhere near — though it still carries billions in cumulative losses and has warned that profitability may not hold across the full year as its spending ramps. OpenAI, by contrast, has told investors it does not expect to be profitable until around 2030, after spending on the order of hundreds of billions of dollars on computing. What they share is that private capital has been willing to fund both stories so far. The question an IPO forces is whether a broader base of investors — pension funds, retail buyers, index trackers — will underwrite them once the numbers, the compute bills and the revenue assumptions are laid out in black and white.

Why it matters

The AI industry has been running on private capital and private valuations — or, as with Alibaba's $10 billion share sale this weekend, on the balance sheets of the giants that already trade publicly. For the pure-play labs, the numbers have so far been whatever the last funding round agreed they were. An IPO ends that. It puts a real, liquid, public price on the frontier — and it ties the fortunes of these labs to the daily verdict of a market that does not care how impressive the technology is if the economics do not eventually work.

Two of the most important companies in the world are about to be priced by everyone, not just by insiders. Whatever the tickers end up being, the AI era is about to get its first real public scoreboard.