Investing
Invesfera Newsroom
Investing

Anthropic's IPO asks you to pay for perfection: treat it as a price test

At $70-$78 a share, Anthropic's debut prices in flawless growth, huge compute bills and benign markets. Size any position as a bet on the price, not a lottery ticket.

Published
November 10, 2026
Reading time
3 min
Categories
Investing

AI-generated image

Is a company worth the price when the price assumes nothing will go wrong? That is the question hanging over Anthropic's Nasdaq debut, and the answer matters more to a retail investor than the thrill of a ticker symbol.

The basics first. Anthropic set an indicative range of $70 to $78 a share for 825 million shares, according to CNBC, and plans to list as "ANT." The offering could raise up to $64.4 billion, plus $26 billion from a private placement with cornerstone investors including Google, Nvidia and Microsoft, Reuters reported. An earlier version of the pricing details, with a higher top end, briefly reached banks and was retracted.

The headline valuation depends on who is counting. The Wall Street Journal's headline says $1.3 trillion. Others cite up to about $1.02 trillion, and CNBC's fully diluted calculation reached $1.405 trillion. The reports don't settle which basis is right, so we won't pretend to a precise figure. Whichever you pick, it is a trillion-dollar-plus price on a company last valued privately at $965 billion.

The growth is real, but the slowdown is in the filing too

Nobody should wave away the business. Third-quarter revenue reached $16.96 billion, up 47% from the prior quarter, and the net loss narrowed. That is remarkable.

But the same report notes that growth had surged 144% in the previous quarter. Going from 144% to 47% is still fast, yet the direction is the point. A valuation this size does not need growth to stay good. It needs growth to stay extraordinary for a long time. In our view, that is what "paying for perfection" means: the price leaves little room for a merely very good outcome.

The bill for that growth is enormous

The company has nearly $900 billion in computing commitments with partners including Amazon and Google, some of it dependent on financing outside its control. We think this is the most underrated number in the story. Revenue is the visible part; the commitments are the fixed obligations behind it. If demand wobbles, the revenue side flexes and the commitment side doesn't.

Add the filing's own warnings: competitors could build models faster, and AI progress could outrun safety and security measures. Companies rarely write risk factors for sport. Read them as the company's own list of what could go wrong.

The market is not in a generous mood

Timing matters. Reuters describes a second half in which investors question whether AI returns will justify the spending and debt costs at elevated interest rates. The Nasdaq just had its deepest one-day drop in 19 months, on fears of more AI scrutiny after the midterms. Some prospective investors were reportedly pushing for a more attractive valuation, viewing a recently discussed $1.5 trillion as the high end of expectations.

That last detail cuts both ways. A range below that talked-about figure suggests the banks listened to buyers, which is sensible. But "less expensive than the rumor" is not the same as "cheap."

How to treat it: a price test, sized like one

We're not telling anyone to buy or avoid the shares, and nothing here is personal advice. What we'd argue is a framework.

First, separate the company from the stock. Anthropic can be a great business and ANT can still be a poor purchase at a given price. Second, expect volatility on day one and beyond. Big, hyped listings tend to be traded on emotion early on; that is an opinion, not a promise. Third, size the position so that a deep drop would hurt your plans very little. If losing half of it would change your life, it is too big.

Fourth, you don't have to be there at the opening bell. The company will report results as a public company, and the market will update its view. Waiting for a first earnings report or two costs you the pop, if there is one, and buys you information. For most savers, that trade looks attractive.

The verdict: this IPO is a genuine test of how much the public will pay for frontier AI after a bruising stretch. Let institutions run that experiment with big cheques. If you take part, do it with a small slice, a written reason and a price at which you'd change your mind. A lottery ticket costs a few dollars. This one costs a trillion-dollar valuation's worth of expectations.

Debate topics

No topics yet: start the first one.

More stories