Anthropic, which is preparing for what could be a historic IPO, is facing a clear contradiction: while the company is meeting prospective investors, its co-founder and CEO Dario Amodei is urging the AI industry to slow the development of its most powerful models.
According to CNBC, the company confidentially filed its IPO paperwork in June and could list its shares as soon as next month. Anthropic was valued at $965 billion earlier this year, and market expectations suggest it could seek as much as a $2 trillion valuation in the offering. CNBC also confirmed that the company has selected Nasdaq for its potential IPO.
In an essay published on Saturday, Amodei proposed a three-step approach that he said would temper how quickly model capabilities improve without sacrificing “commercial advantage” or the United States’ lead in AI. His plan calls for model developers to open themselves to independent third-party evaluators, for frontier companies to establish “common safety standards,” and for democratic countries to coordinate with authoritarian governments “to the extent this is possible.”
The proposal arrives as the public and industry debate over AI risk grows more intense. In recent weeks, several researchers have warned that advanced models could pose catastrophic dangers. Against that backdrop, United States President Donald Trump attacked Amodei on Truth Social, arguing that AI does not need extra “guardrails” but rather a “STRONG AND SMART” president.
At the same time, OpenAI CEO Sam Altman publicly backed Amodei’s proposal. Elon Musk also expressed support; he leads SpaceX, which owns Grok creator xAI. Still, Altman told Fortune that “Right now would be an ill-advised moment to go public,” reiterating that OpenAI is not targeting an IPO until next year. The company’s finance chief Sarah Friar told employees that the AI lab “will be a public company in 2027.”
Market opinion is divided. D.A. Davidson analyst Gil Luria said investors may not necessarily see the slowdown push as a negative as long as companies are not actually walking away from IPO plans, additional compute or further model training. Lise Buyer, a partner at IPO advisory firm Class V Group, said growing fears around AI may not alter IPO timing, but they could affect valuation as the investment case becomes more tightly linked to questions of control and risk.
Anthropic’s financial performance remains strong. CNBC previously reported that the company reached $65 billion in annualized revenue in July, about seven times higher than a year earlier. Two sources also told CNBC that Anthropic has informed some shareholders it expects to post operating profit for a second consecutive quarter in the current period. The Financial Times reported that development on Sunday. Menlo Ventures partner and Anthropic investor Matt Murphy described the growth rate as “off the charts” and said a public listing would bring greater transparency to the business.
Valuation, accountability and market risk
Some investors and analysts argue that an IPO could actually help Anthropic by adding transparency, scrutiny and accountability. Altimeter Capital CEO Brad Gerstner, whose firm invests in both Anthropic and OpenAI, wrote on X that bringing more “transparency, scrutiny, accountability” and broader participation to AI companies is “crucial.” He added that the market knows how to price risk, pointing to SpaceX as an example.
Critics, however, say stricter safety standards may end up benefiting the biggest players. Gartner analyst Arun Chandrasekaran said that could favor Anthropic and OpenAI if smaller rivals cannot afford the safety, evaluation and security investments required for frontier-level models. Gil Luria went further, saying the two companies may be engaging in “monopolistic behavior.” Wired reported that OpenAI has even asked members of Congress for guidance on whether a coordinated industry slowdown could violate antitrust law.
The debate matters beyond those two companies. Anthropic has signed a series of multibillion-dollar compute deals this year with Nscale, Advanced Micro Devices, SpaceX and Google. OpenAI told investors in February that it is targeting roughly $600 billion in total compute spend by 2030. Both companies are major users of Nvidia graphics processing units, meaning any change in the pace of AI development could ripple through infrastructure spending and across major tech suppliers.
PitchBook analyst Harrison Rolfes warned that model-company valuations may now deserve a discount because investors still struggle to trust that the technology can be commercialized safely. Deepwater Asset Management managing partner Gene Munster said the market is currently underwriting “exponential uninterrupted improvements” in models, so any perceived slowdown would be negative. Even so, he expects the AI leapfrog race to continue because the long-term opportunity is too large for leading companies to genuinely slow down.
