When we wrote about Anthropic’s impending initial public offering last month, one reader called it a classic top-of-the-market debut that would mark the peak of the artificial-intelligence trade. Unfortunately for Anthropic, the peak—or, at least, a peak—may have already passed.

Anthropic attempted to woo Wall Street at a presentation last week, but slowing quarter-over-quarter sales growth gave some investors pause. The company is now expected to go public in November, with some investors discussing a valuation of around $1.5 trillion, down from the roughly $2 trillion previously expected, according to Reuters.

Adding to the scrutiny, The Wall Street Journal reported Tuesday that Anthropic lost $80.7 billion in the first half of 2026, citing a draft S-1 prospectus. More than $70 billion of that loss came from accounting charges tied to financing instruments the company had issued.

Anthropic is also on the hook for more than $600 billion in AI infrastructure costs, the Journal reported, with negotiations underway that could push that number to roughly $750 billion. Those are staggering figures at a time when investors are growing wary of a potential AI bubble.

Anthropic declined to comment.

Against that backdrop, Anthropic’s debut is no longer shaping up as a crowning celebration for the AI revolution. Instead, the offering—starting with the public release of the prospectus—is becoming a major test for the AI trade.

“The tenor of conversations with clients this week indicated increasing skepticism that the S-1 and IPO could be the definitive catalyst to restart the AI trade,” Dauvin Peterson, an analyst at 22V research, wrote in a note Sunday. “Investors are taking a much more critical look at the business and returns.”

The “burden of proof” is now on Anthropic to make the case for why its offering is a good investment, Peterson says. That is a departure from a few months ago, when finding an Anthropic skeptic on Wall Street was like searching for a unicorn.

Peterson says the company needs to offer revenue estimates through the end of 2027, give more detail on its costs and margins, and address concerns over the falling price of tokens—units of data processed by AI models.

It doesn’t help that other AI IPOs are faltering at the exact same time.

On Tuesday, South Korean cloud provider Elice’s shares plunged 32% in their Seoul trading debut. It was the second notable IPO setback in as many weeks: Australian cloud company Firmus Grid withdrew its offering after failing to fetch a $30 billion valuation.

The Elice and Firmus setbacks indicate investors are getting more discerning about AI valuations, says Avery Marquez, director of investment strategies at Renaissance Capital, a provider of pre-IPO research and IPO-focused investment funds.

“I think investors will push back on valuation,” Marquez says of Anthropic. “But given that this is a high-profile name and would be the first generative AI pureplay to go public, I think there will be a lot of demand regardless of where AI sentiment falls at that point in time.”

The next question is how the shares will perform after the debut. Once Anthropic is public, its continued success in the market will come down to how the underwriters price the IPO and what kind of expectations investors have, Marquez says.

The recent setbacks suggest investors will have little patience for disappointments.