Anthropic’s record $81.3 billion IPO triggers public doubts over safety and valuation bubble concerns

Global Times
1795171617000

File photo VCG

File photo: VCG

US artificial intelligence (AI) firm Anthropic began trading on the Nasdaq Stock Market on Thursday (US time), with its IPO raising $81.3 billion in what was the largest stock-market debut on record. While some have applauded the mega-IPO, there are also concerns about a valuation bubble amid the rising popularity of China's open-source models.

The company behind Claude Code priced its shares at $83 each, above its initial range of $70 to $78, giving the San Francisco-based firm a market capitalization of around $1.5 trillion ahead of its Nasdaq debut on Thursday. Anthropic's IPO eclipses the record-setting debut of Elon Musk's SpaceX in June, Reuters reported.

Yet on the very day Anthropic priced its IPO, dozens of top AI researchers, startup founders and scientists from across Silicon Valley co-signed an open letter - The MisAnthropic Statement - openly declaring their opposition to the company.

"The company uses safety branding to court investors, employees, and the public - even as it has quietly weakened its safety pledges and lobbied for self-serving regulation. Anthropic is not a safe AI company, and we must reject its reckless attempt to amass power in the name of the greater good," the letter stressed.

In a report in October, independent financial research provider New Constructs called Anthropic's upcoming offering the "most ridiculous IPO of 2026," CNBC reported.

Anthropic's mounting operating losses coupled with emerging competition from a plethora of open-source models led New Constructs to conclude that, "We don't think Anthropic has a viable business."

"Since the arrival of open-source models, it's been clear that the closed models would struggle to generate profits," the firm wrote.

"From a trading perspective, US tech stocks need a major IPO like Anthropic's to shore up confidence. However, the AI firm's valuation has fallen from an initially estimated $2 trillion to $1.5 trillion after the startup reportedly disclosed financial details at a recent investor event. That signals that the market is reassessing how it values the AI growth story," Xiang Ligang, president of the Beijing Institute of Intelligent Era Industry, told the Global Times on Friday.

Looking at the industry itself, a company built primarily on training AI-powered models cannot justify a valuation of this magnitude on its pitch to institutional investors and brute-force computing power alone. Ultimately, it has to prove that it can reach consumers and turn its model capabilities into recurring, sustainable revenue - a market already crowded with competitors pursuing similar strategies, Xiang said.

A report released by Silicon Valley venture capital firm Menlo Ventures earlier in November showed that Chinese AI models, led by Moonshot AI's Kimi series, now account for 89 percent of estimated global enterprise spending on open-source systems, according to media reports.

Anthropic has committed to nearly $900 billion in computing spending with infrastructure partners, including Amazon and Alphabet's Google, to train and operate AI models, CNBC reported, citing the company's prospectus.

It warned that its computing providers and infrastructure partners depend on third-party financing and credit support to acquire chips and build capacity, per Reuters.

"If such financing or support becomes unavailable or more ⁠costly, contracted capacity may be delayed, repriced, reduced, or not delivered," Anthropic said in an IPO filing.

Long-term compute commitments shift the focus from valuation to the sustainability of the business model: capacity and costs are locked in, while revenue remains variable. Many enterprise customers have not signed multi-year contracts and are billed based on actual AI usage, allowing them to dial consumption up or down as their needs shift. This puts Anthropic's ability to turn committed compute capacity into recurring cash flow to the test, Xiang said.