Anthropic's Two Trillion Dollar IPO Rests on a Revenue Number From 2028

Anthropic is preparing what would be the largest public offering ever attempted, and bankers are reportedly circling a valuation approaching 2 trillion dollars. The figure does not rest on what the company earns now. It rests on a forecast of 190 to 200 billion dollars in revenue in 2028, two years out from a business that was running at 47 billion as recently as May.

Most mature companies get valued on earnings, or on some version of them. Anthropic is being valued on a multiple of revenue it has not made yet, in a year that has not happened. That is unusual but not unprecedented, and the precedent is recent: Cerebras and SpaceX both went public on forward looking methodologies rather than trailing numbers. What makes this one different is the size of the gap being bridged and the size of the company on the other side of it.

The near term numbers are the part of the pitch that is checkable. Anthropic expects at least 10.9 billion dollars of revenue in the second quarter, more than double the quarter before it, and a first quarterly operating profit of 559 million. That is the argument in miniature. Enormous infrastructure spending, revenue compounding faster than the spending, and margins that are supposed to open up as the thing scales. The comparables Wall Street is reaching for are Palantir at 53 times expected 2026 revenue, and Cloudflare and SpaceX at 41.6 times each.

Markets took the news well and unevenly. Semiconductor shares rose on Monday even as the S&P 500 fell 0.52 percent to 7,744.97, with Microsoft down 3 percent and Meta down 3.5 percent on rising yields. Tokyo followed overnight, Advantest up 3.9 percent, Kioxia 2.9 percent, Tokyo Electron 1.6 percent. Whatever investors think of the valuation, they are clear about who gets paid while it is being built.

The skepticism is not hidden. A principal at Aleph Investments said the company could probably get to 2 trillion and then wondered aloud whether it would stay there. That is the whole question. A forward multiple prices a future, and the market gets to remark it every quarter afterwards, which is a different kind of scrutiny than a private round. There is also the small matter of the 30 year Treasury yield hitting its highest level since 2007 on the same day, partly because of corporate borrowing to fund AI infrastructure.

So the biggest listing in history is being underwritten by a spreadsheet that runs to 2028. It might well be right. But the number that makes it work is still two years away, and the market will have eight quarters to change its mind about it before then.