The valuation on the Anthropic IPO has collapsed from nearly $2 trillion toward $1.2 trillion in a matter of weeks. As the roadshow begins, Wall Street is no longer asking how high the deal can go. It is asking how low it has to fall to find buyers.

By Blumefield | November 9, 2026

The Haircut Begins

Wall Street is slashing the Anthropic IPO before a single share has changed hands. In the space of three weeks, the number investors are willing to pin on the maker of Claude has fallen from close to $2 trillion toward a range that now starts with a one, and the bankers running the deal are feeling the pressure in real time.

As the roadshow prepares to hit the road this week, with a public debut targeted for the days before Thanksgiving, prospective buyers have been guiding Anthropic and its advisers toward a valuation closer to $1.2 trillion to $1.5 trillion. That is a staggering reset from the roughly $2 trillion figure floated when the company began briefing investors in October. The Anthropic IPO was supposed to be the coronation of the frontier AI era. Instead it has become its first public stress test.

The stakes are hard to overstate. This is the first time a pure-play frontier lab will ask ordinary public markets to put a price on the economics of artificial intelligence, rather than leaving that job to a closed circle of venture firms and sovereign wealth funds. Every institution weighing an allocation knows that the number they settle on will echo through OpenAI, xAI, and the hyperscalers that have bet hundreds of billions on the same thesis. What looked like a formality in September now reads like a referendum.

A Revenue Rocket With a Loss Problem

The case for the bulls is genuinely spectacular, and it is worth stating plainly before the caveats arrive. According to figures Anthropic presented at its October investor day and in its registration filing with the U.S. Securities and Exchange Commission, third-quarter revenue reached roughly $17 billion, up about 45 percent from the previous quarter. The company told investors it was profitable on an adjusted basis during the period. For a business that booked around $4.6 billion in revenue for all of last year, a more than twelvefold jump, this is one of the fastest climbs the technology industry has ever recorded.

Then comes the other column of the ledger. The same filing that showcased the growth also disclosed a loss of roughly $42 billion in 2025, the bulk of it tied to non-cash charges, with an operating loss north of $8 billion once the writedowns are stripped out. Revenue growth, while enormous in absolute terms, is decelerating in percentage terms. The sequential jump from the second quarter to the third was smaller than the near doubling that came before it. That deceleration, modest as it sounds, is exactly the kind of signal public investors seize on when they are being asked to pay a premium measured in trillions.

The tension at the heart of the Anthropic IPO is this collision between a revenue line that behaves like a rocket and a cost base that behaves like a utility building out a national grid. Private backers were content to underwrite the gap on faith and access. The investors now circling the Anthropic IPO are a different breed, and they are doing the arithmetic in public.

The $895 Billion Question

Nothing captures the scale of the bet better than Anthropic's compute obligations. The company has lined up close to $895 billion in computing purchase commitments and data-center leases to feed its models, with a large share of that spending due before the end of the decade. Those contracts are the fuel for the growth story. They are also an anchor chained to the balance sheet.

This is the question institutional buyers keep returning to. To justify even a trimmed valuation, Anthropic has to convert astronomical compute spending into durable, high-margin revenue faster than its own price cuts and competition erode it. The filing noted that roughly $600 million in price reductions partly offset revenue in the latest quarter, a reminder that the market for frontier intelligence is already a price war as much as a technology race. Cheaper tokens can mean bigger bills for customers and thinner economics for the provider at the same time.

Anthropic intends to list on the Nasdaq under the ticker ANT, and its leadership, including chief executive Dario Amodei and chief financial officer Krishna Rao, has reportedly signaled a preference for pricing the deal conservatively to leave room for a first-day pop. That instinct tells you how the mood has shifted. When demand is feverish, issuers push valuations as high as the market will bear. When demand is nervous, they price for a bounce and hope momentum does the rest. You can read Anthropic's full case on the company's own investor materials, but the market, for now, is writing its own version.

Why Wall Street Blinked

The timing could hardly be worse, and that is no accident. The Anthropic IPO is walking into a market that has spent the past fortnight questioning the entire AI trade. A brutal selloff in late October and early November wiped out hundreds of billions in combined market value across the largest technology names, dragged down by disappointing guidance from chipmakers and renewed fears that capital spending has outrun any plausible return. Intel and Alphabet both posted double-digit declines on the back of their results, and the neocloud providers that lease out graphics processors saw their shares and bonds punished in tandem.

Against that backdrop, the concerns investors are voicing about the Anthropic IPO are less about the company than about the category. Several cite overexposure to AI across their existing portfolios, having already loaded up on Nvidia, the hyperscalers, and a long tail of infrastructure plays. Others point to a political dimension that barely registered a year ago, as voters in dozens of local ballots have begun rejecting the data centers that AI demands, raising the cost and the uncertainty of the buildout. Underneath it all sits a nagging doubt about whether giants like Amazon and Meta will ever earn an adequate return on the sums they are pouring into the technology.

Those doubts do not negate Anthropic's growth. They do change the multiple a disciplined investor is willing to apply to it. A valuation that looked visionary when the Nasdaq was making new highs looks aggressive when the index is shedding a trillion dollars in a week. Markets price risk, not just promise, and the risk premium on everything with "AI" attached has quietly repriced higher.

The First Domino

What makes the Anthropic IPO so consequential is that it will not stay contained. A clean debut, priced sensibly and trading up, would hand the entire sector a template and embolden OpenAI, which is widely expected to pursue its own listing in the coming year. A sloppy debut, or a last-minute decision to delay, would send the opposite message and could freeze a pipeline of AI offerings that bankers have been counting on to define the next two years of capital markets.

There is a plausible path to a good outcome. Some portfolio managers argue that even a $1.2 trillion price should find willing buyers, given Anthropic's growth rate and its position as one of only a handful of credible frontier labs. Price the deal conservatively, let scarcity do its work, and a debut that opens near $1.2 trillion could drift back toward $2 trillion as momentum returns. That is the bull case, and it is not unreasonable.

Yet the very fact that the conversation has moved from how high to how low in a matter of weeks is the story. The Anthropic IPO began as a victory lap and has turned into a negotiation, and negotiations reveal who holds the leverage. For most of the AI boom, the labs have held it, dictating terms to investors desperate for access. The repricing of the Anthropic IPO is the first clear sign that the public market intends to hold some leverage of its own. Blumefield will be tracking the roadshow and the final pricing closely, and readers can follow our continuing coverage of the AI markets at Blumefield. For context on how Anthropic reached this moment, see our earlier reporting on the company's landmark filing at Blumefield.

Whatever number the bankers settle on in the days before Thanksgiving, the lesson is already clear. The era in which artificial intelligence could name its own price has met the one force that has humbled every boom before it, which is a buyer who can simply say no.