The Anthropic IPO Was Supposed to Save AI. Then the Market Pulled the Rug.
The IPO Was Supposed to Save AI. Then the Market Pulled the Rug.
Anthropic’s public debut was supposed to be a turning point. Instead, as the company’s shares climbed, the broader tech market sank. The contrast says a lot about where investor confidence stands heading into the end of 2026.
The Clearing Event That Didn’t Clear
Investors had hoped the Anthropic IPO would act as a “clearing event,” lifting tech after months of choppy trading, the Nvidia espionage scandal, and growing fears that AI chip sales to Asia would slump under tighter U.S. export controls. As reported in CNBC, the day went the other way. The Nasdaq Composite closed down 2.25%, and the S&P 500 fell 2% to 7,175.
The company itself did well. Its shares closed up 17% from the IPO price, and the offering raised more than $81 billion, the largest IPO in U.S. history.
“There’s a reset of AI in general.” — JJ Kinahan, Cboe Global Markets
Export Jitters and a Short Seller’s Report
Several old worries returned at once. As noted in CNBC, the short seller Hunterbrook Media alleged that firmware sold by AMD had turned up in U.S.-blacklisted companies. AMD stock fell more than 7%, though the chipmaker told CNBC it complies with U.S. law. Separately, documents released by the House Foreign Affairs Committee showed that a Commerce Department subagency had approved the sale of more than 879,000 AI chips to Chinese companies. That raised the prospect of a slowdown in Asian chip sales as Washington tightens controls.
A Price War Nobody Wins
The deeper problem, according to investor Dan Niles, founder of Niles Investment Management, is competition between Anthropic and OpenAI. As CNBC reported, Niles said the price war between the two frontier labs has hurt both, eroding the revenue and profit potential that justifies the enormous spending on AI infrastructure.
His numbers make the case. Combined revenue growth for OpenAI and Anthropic slowed from 75% sequentially in the second quarter to 48% in the third. Their combined losses widened from $7.9 billion to about $8.4 billion over the same period. Niles described himself as a “circumspect bull,” comparing the moment to 2022, when the major indices sank in the second half of the year. “You could have still made money that year,” he said. “You just needed to have the right combination of longs and shorts.”
“You can’t buy things blindly and they’ll all go up.” — JJ Kinahan, paraphrasing a shift in investor attitudes
Buy-the-Dip Loses Its Magic
For years, tech investors have made money by buying every dip. Kinahan says that habit is finally breaking down. Nvidia’s ongoing declines have shaken confidence in the strategy, and the mood now is more selective. Riskier corners of the market took the heaviest hits. CoreWeave, one of the highly indebted “neoclouds” that provide dedicated AI computing, closed down more than 6%. Hyperscalers such as Alphabet and Amazon, usually seen as the more defensive AI bets, also finished lower.
What Comes Next
Anthropic’s next few trading sessions will serve as a gauge of market strength. Kinahan called Friday a potential “wild card” that could mark a “final flush out” of selling, or a bounce that leaves the market in “no man’s land.” The bigger test, he said, comes Monday, when options on Anthropic begin trading. Retail investors, who tend to buy calls, will reveal how much faith they have in the stock and in the AI trade more broadly.
There was one bright spot. Cybersecurity stocks kept rallying, with Palo Alto Networks and CrowdStrike both closing higher. AI is uncovering vulnerabilities across software infrastructure faster than they can be patched, and that scramble is creating demand of its own.
For now, the verdict on the AI trade is still out. The IPO proved that investors will pay enormous sums for a piece of the future. The week also showed that they are no longer willing to buy everything that carries the AI label.
Aaron Rose is the founder and publisher of Tech Reader Magazine, covering system architecture, cloud platforms, and AI policy.