Anthropic's Tumble Shows AI IPO Fever Ignores Burn Rate
Anthropic shares crashed 8.5% on their second day, nearly touching the $83 offer price. The sell-off exposes a wider truth about AI valuations: growth rates don't compensate for questionable unit economics.
The Offering Price Was the Real Target All Along
Anthropic opened at $88.91 on Tuesday, down 8.5 percent from Monday’s close — and intraday it touched $83.05, within pennies of the $83 offer price. The market cap slipped to roughly $11.6 billion, ejecting the company from the top ten in U.S. public listings, behind Berkshire Hathaway and Eli Lilly.
Monday’s 17.1 percent pop looked like a triumph. Tuesday looked like a correction with teeth. What happened between those two days matters more than either number alone.
The intraday dip to $83.05 was significant not because the stock collapsed, but because it proved how quickly enthusiasm evaporates when earnings visibility is thin. In the first twelve minutes of trading, roughly $420 million in shares changed hands, with sell orders outnumbering buys by nearly three to one according to order-flow data reviewed by Bloomberg. By midday, the stock had stabilized around $87, but the damage to sentiment was already priced in.
This is not unusual for a hot IPO. What is notable is how quickly the regret set in. One trading day.
Who Bought, and Who Regrets It Now
The $11.95 billion raised across the $813 million public offering, $26 billion in private placements, and a fully exercised greenshoe option was the largest AI-related capital event in history. The institutional books are full. And they’re already underwater.
Emmanuel Balavarinos at Forte Securities told Bloomberg that investors are on a buying strike across AI and technology broadly. This isn’t just an Anthropic problem. It’s a category problem.
Scott Report downgraded the stock to neutral, noting the valuation can only be justified if revenue growth accelerates sharply over the next two years. Anne Milletti at Allspring Global Investments put it plainly: the question on every investor’s mind right now is “where does the profit come from?”
That is not a speculative question. It is a fundamental one.
But the pressure extends beyond balance sheets into portfolio strategy. Several large-cap mutual funds with mandated tech weightings found themselves forced buyers at the October 1 close, creating a structural floor that may delay further declines but also locks in unrealized losses through earning season. When those quarterly reports arrive, the gap between growth narrative and margin reality will be impossible to paper over.
The Moat Question Anthropic Can’t Answer Yet
Anthropic’s pitch has always been safety-first AI. Claude is positioned as the responsible alternative to open-weight models and the more aggressive product cycles of competitors. That narrative won praise and venture capital. It did not win a pricing advantage on day two of public trading.
Compounding the pressure was news that Apple is building its own AI agent to compete with Claude. Apple does not need a third-party model. When the biggest distributor of AI capabilities decides to build in-house, every outside vendor gets re-evaluated.
A moat made of alignment research papers does not block a silicon valley ramp. Claude may be safer, but safety is not distribution. Distribution is scale, integration, and lock-in. Those are things Anthropic trades on, but cannot yet prove it owns.
The second-order effect here is significant. If Anthropic cannot demonstrate pricing power or enterprise stickiness, the same dynamic plays out across the sector. Cohere, Mistral, Perplexity — all companies positioning as differentiated API providers rather than infrastructure owners face the identical threat. An Apple, Google, or Amazon moving deeper into proprietary agents doesn’t just compete for users. It removes the platform dependency that third-party model vendors rely on for leverage.
The Timing Was Suspect, Even by Silicon Valley Standards
Byron Deeter at Benchmark, an early backer, told CNBC that almost no company would have attempted this listing under current conditions. The macro backdrop includes post-midterm sentiment headwinds, data-center capacity concerns, Nvidia CEO Jensen Huang’s recent stumble, and Federal Reserve rate-hike risk.
Deeter praised the team for holding course and closing the record fundraise anyway. But the fact that the deal closed at all in this environment says something about internal urgency — or external commitment. Either way, the market did not care about the story. It cared about the multiple.
What Deeter did not say aloud — and analysts watching closely are noting — is that the timing reveals something about venture capital’s posture toward public markets. Firms that staged substantial secondary sales alongside this offering effectively locked in gains at peak sentiment. For limited partners who invested at Series funding rounds priced well below $83 per share, the exit window mattered more than anything that happens at quarter end.
What Happens Next
The immediate trajectory is clear. Shares traded below the offer price would trigger seller distress among any retail buyer who caught the Monday pop. Short interest could expand if institutions begin reducing positions rather than averaging down.
The longer arc depends on whether Anthropic can demonstrate revenue growth that justifies its valuation relative to peers, or whether the market continues repricing AI infrastructure plays downward. Scott Report’s two-year timeline is a useful benchmark. If Claude usage and enterprise contracts do not scale aggressively by mid-2028, the neutral rating becomes a sell recommendation by default.
But the broader lesson cuts deeper than one company’s trajectory. The market once rewarded narrative in AI. It now demands unit economics. Anthropic raised more money than any AI company before it. It will now be held to harder standards than any of them. That is not a punishment. It is the natural consequence of a category that priced itself into a corner.
For every analyst warning about burn rates, there is a competitor still spending freely to buy attention. OpenAI’s ChatGPT Pro tier launch, Google’s Gemini enterprise push, and Microsoft’s continued Copilot bundling all signal that the race is far from settled. The difference now is that Anthropic will have to fund its side of that race from public-market capital, where scrutiny is mandatory and patience is scarce.
The tumble is not a verdict. It is a recalibration. And in markets that forget what they paid too quickly, the reminder arrives sooner rather than later.