Anthropic shares slide as AI selloff tests record IPO debut

A steep premarket drop on Friday signals investor caution around AI valuations, one day after the biggest IPO on record

Anthropic shares slide as AI selloff tests record IPO debut

Anthropic, the maker of the Claude artificial intelligence assistant, fell more than six per cent in premarket trading on Friday, one day after completing the largest initial public offering in history - a rough start that underscored deepening investor uncertainty about whether the economics of the AI sector can justify its lofty valuations.

Shares in the San Francisco-based company were down approximately 6.4 per cent at US$90.94 in early Friday trading, according to CNBC, after closing their first session on Thursday at US$97.17 - itself a 17 per cent gain from the IPO price of US$83. Nvidia fell roughly 2.7 per cent in the same premarket window, while Advanced Micro Devices dropped approximately 5.4 per cent.

The selloff arrived even as Anthropic's market debut was being celebrated on Wall Street. The company raised US$81 billion in its listing on the Nasdaq under the ticker ANT, surpassing the record set by SpaceX's June 2026 IPO. At its Thursday close, Anthropic carried a market value of approximately US$1.3 trillion, or around US$1.8 trillion on a fully diluted basis, according to The Wall Street Journal.

For Canadian financial advisors, the two-day swings in Anthropic shares crystallise a question that has shadowed the AI trade throughout 2026: whether the sector's most prominent names are priced for a future that remains uncertain. As Wealth Professional reported ahead of the offering, Canadian investors have been watching Anthropic's debut closely even though the company lists only on the Nasdaq. Matt Kennedy, senior strategist at Renaissance Capital, told Reuters in connection with that report that a successful IPO from Anthropic could "crack open or slam shut" the IPO market well into 2027.

What drove Friday's drop

Friday's premarket losses were not attributable to any Anthropic-specific development. The drag came from broader sector pressure. A short-seller report from Hunterbrook Media on Thursday alleged that firmware sold by AMD had appeared in US-blacklisted companies; AMD closed down more than seven per cent on Thursday, according to CNBC. The chipmaker told CNBC it complies with US law.

Separately, the House Foreign Affairs Committee released documents showing that a subagency of the US Commerce Department approved the sale of more than 879,000 AI chips to Chinese companies - a disclosure that raised concerns about a potential sales slowdown under tightening export controls. Nvidia has also been navigating turbulence following the resignation of longtime chief executive Jensen Huang.

The broader market offered little support. On Thursday, the Nasdaq composite dropped 2.3 per cent, the S&P 500 fell two per cent for its sixth consecutive losing session, and the Dow Jones Industrial Average declined 2.2 per cent, according to The Wall Street Journal.

Capex questions overshadow the listing

Underlying the market's caution is a structural concern that extends well beyond Anthropic's debut week.

In its 2027 Outlook report released this month, Morningstar Wealth warned that AI valuations rest on a "build-it-and-they-will-come assumption" - that today's infrastructure spending will be repaid by tomorrow's demand.

Morningstar estimates that the leading hyperscalers - Alphabet, Amazon, Microsoft, and Oracle - collectively committed approximately US$750 billion to capital expenditure in 2026, with spending expected to approach US$1 trillion in 2027.

By contrast, Morningstar estimates that leading AI labs including Anthropic and OpenAI generated around US$100 billion in combined revenue during the same period. That gap, Morningstar argues, sets a high bar: for investments in AI compute to earn attractive returns, AI companies must turn technological progress into commercially valuable products and retain enough of the resulting profits - an outcome the report describes as "far from assured."

Morningstar's report also flagged a risk specific to the structure of the AI industry itself: the ownership and revenue relationships between hyperscalers, AI labs, and chipmakers are deeply intertwined. Amazon and Alphabet hold meaningful stakes in Anthropic while also serving as its primary cloud providers. Nvidia has invested in Anthropic, OpenAI, and CoreWeave - the same companies that drive demand for its chips. That circularity, Morningstar argues, means a stumble at any one point in the chain could affect earnings, cloud backlogs, and suppliers' orders simultaneously.

What it means for advisors

The week's events reinforce the case for scrutinising AI holdings against underlying fundamentals rather than momentum - a discipline that becomes especially important when clients are drawn to high-profile names. As Wealth Professional has reported, the best advisory teams in Canada recognized in the 2026 5-Star Advisory Teams report are building practices around rigorous risk management and client communication during volatile periods - skills that will be tested as the AI trade continues to swing.

Renos Savvides, head of equity capital markets at Neuberger, told The Wall Street Journal that the two-day performance left the broader question open. "By any means, Anthropic's IPO was a success," Savvides said. "But this was looked at as a barometer for the AI trade, for the IPO market and for the broader market itself. By that measure, the jury is still out."

For more context on how the AI buildout is reshaping equity markets, advisors can read Wealth Professional's earlier coverage of what to expect from the Anthropic IPO, which examined the company's business fundamentals and the challenges its valuation presents for long-term investors.

LATEST NEWS