Anthropic is about to discover whether private-market enthusiasm for artificial intelligence survives contact with Wall Street. There is little reason to doubt demand for AI itself. Businesses are spending aggressively on models, computing power and the infrastructure required to run them. The harder question is whether enormous demand for AI necessarily makes the companies supplying it good investments.

Anthropic has built one of the fastest-growing technology businesses in the world, but doing so has required extraordinary amounts of capital. Compute is expensive, remaining at the frontier is becoming even more expensive, and sustained profitability is still some distance away. Investors are therefore being asked to place a price today on cash flows that may only become reliably positive much further down the road. AI may be revolutionary, but revolutions still have to pay their bills.

The timing makes the IPO even more interesting. Anthropic is preparing to enter public markets just as technology stocks are becoming less forgiving of disappointing numbers. Its prospectus offers investors one of the clearest looks yet at the economics hidden beneath the AI boom.

Anthropic may therefore represent more than another technology listing. It could be judgement day for AI. If Wall Street is willing to look beyond the losses, financing needs and enormous infrastructure commitments, the party continues. If not, investors may finally ask Silicon Valley the question it has spent years avoiding: when does all of this actually make money?

Behind the Hype: What Anthropic’s Prospectus Really Says

Anthropic’s prospectus does not exactly ease investors gently into the numbers. Revenue increased from just $41 million in 2023 to $4.6 billion in 2025, before reaching $16.3 billion in the first half of 2026 alone. Second-quarter revenue was up 1,368% year-on-year1. If Wall Street wanted evidence that businesses are willing to pay for artificial intelligence, Anthropic has certainly provided it.

The problem begins when you look at what it costs to stay at the frontier. Compute and infrastructure expenses reached $8.9 billion in the first six months of 2026, almost four times their level a year earlier. Anthropic still reported an operating loss of $693 million, although that is a considerable improvement from the $2.1 billion lost operationally during the first half of 2025. The business is therefore moving towards profitability remarkably quickly, but it is doing so while feeding an infrastructure machine that demands increasingly enormous amounts of capital.

Then comes the headline number: an $80.7 billion net loss for the first half of 2026. That figure looks terrifying, but investors should resist throwing Claude out with the bathwater. Roughly $72.4 billion came from changes in the fair value of convertible notes, derivatives and SAFEs, and a further $7.6 billion from other non-operating items, rather than the day-to-day operation of the business.¹ The $693 million operating loss is therefore a much better indication of how the underlying company is performing.

That does not make the prospectus comfortable reading. Anthropic has already raised $141.1 billion of capital, held more than $130 billion across cash and short-term investments at June, and has committed more than $600 billion to future computing infrastructure2. The awkward mismatch is that much of its revenue remains consumption-based while many of those compute obligations extend years into the future. Anthropic itself warns that slower revenue growth could leave it paying for infrastructure that fails to generate the expected returns.¹

And slower growth will come. Even Anthropic explicitly tells investors not to expect current growth rates to continue. The market will therefore have to decide how much it is willing to pay for a company whose extraordinary valuation rests not on whether demand exists today, but on how much of that demand survives once AI stops being the newest toy in the corporate budget.

That is what makes this IPO so important. Anthropic will give public markets a price for something private investors have spent years valuing largely behind closed doors. A successful listing would provide another justification for the enormous valuations sitting across the AI ecosystem. A disappointing one would force investors to ask a much nastier question: if one of the strongest revenue stories in artificial intelligence cannot justify the price, who can?

When Wall Street Has to Price the Dream

Anthropic could hardly have chosen a more interesting moment to ask public markets for their verdict. The Nasdaq fell 2.6% on 22nd October, its worst session since June, as investors once again questioned whether the enormous sums being poured into artificial intelligence will ever generate sufficient returns3. Anthropic is therefore arriving just as Wall Street appears to be rediscovering a concept Silicon Valley has been able to ignore for much of the AI boom: price matters.

The first scenario is the one the industry desperately wants. A strong IPO, particularly anywhere near the $1.5 trillion valuation reportedly being discussed, would suggest investors remain willing to look beyond current losses and value Anthropic on the profits its technology might eventually produce. That would not only validate Anthropic. It could provide another reference point for OpenAI, AI infrastructure companies and the enormous capital expenditure programmes already supporting Nvidia, Broadcom and the wider semiconductor trade.

The alternative is considerably less comfortable. Anthropic continues to operate at a net loss while committing itself to increasingly expensive computing infrastructure, much of which is being supported through complex debt and lease financing. A new $63 billion financing package has already been assembled to fund Broadcom chips destined for Anthropic, including $43.5 billion of senior loans. Demand for Claude may be enormous, but demand does not automatically produce free cash flow.

That distinction matters in a market already becoming less patient. Other AI-linked companies have postponed or withdrawn IPO plans as higher borrowing costs and doubts over AI spending have weakened investor appetite.⁴ If Anthropic struggles, the market may not treat it as an isolated disappointment. Investors could begin questioning the valuations of the chipmakers, data-centre operators and hyperscalers whose own growth assumptions depend upon companies such as Anthropic continuing to spend.

This is why I struggle to see the IPO removing uncertainty from the AI trade. Anthropic can prove that customers want intelligence. What it has yet to prove is that selling that intelligence consistently produces more cash than building it consumes. Until that changes, Wall Street is not valuing a finished business model. It is financing an extraordinarily expensive promise.

Judgement day will not arrive when Anthropic rings the Nasdaq bell. It will arrive when investors stop asking how quickly Claude can improve and start asking when Claude finally pays for itself.

Footnotes

  1. Anthropic, Form S-1 Registration Statement (opens in a new tab), 26th October 2026.

  2. Reuters, Anthropic's soaring sales, steep losses put AI frenzy to IPO test (opens in a new tab), 26th October 2026.

  3. Reuters, S&P 500, Nasdaq end sharply lower as revived AI doubts ignite tech selloff (opens in a new tab), 22nd October 2026.