Anthropic challenged on its safety stance ahead of Nasdaq debut
The publisher of Claude faces a revolt from 89 signatories accusing it of exploiting its AI safety image. The debate erupts as its Nasdaq listing is set to become a test for the entire industry.
Anthropic, the publisher of Claude, is facing a major public accusation just as its Nasdaq debut approaches. On November 18, a coalition of researchers, founders, and legal experts published the "MisAnthropic Statement," asserting that the company is "not a safe AI company." Meanwhile, Anthropic set its IPO price at 83 dollars per share, with first trades expected under the ticker ANT on November 19. The stakes go beyond a single startup turned giant: the entire credibility of the discourse on so-called responsible AI is under pressure.
Une fronde très large contre le discours de sécurité d’Anthropic
The opposition does not come from a homogeneous camp. Among the signatories are Steven Pinker, Max Tegmark, Igor Babuschkin, Thomas Wolf, George Hotz, Gary Marcus, Oliver Hart, Judea Pearl, Laurence Tribe, Lawrence Lessig, and several other well-known names from research and tech. According to the statement, Anthropic uses its reputation for caution to serve commercial and political goals, while pushing increasingly powerful AI systems.
The core of the accusation lies in a contradiction: the company publicly warns of risks related to AI capable of self-improvement, yet continues to pursue that very target. The signatories also reproach Anthropic for weakening its safety commitments and pressuring regulation in ways favorable to its interests. The tone is direct, almost surgical. Not exactly the kind of greeting card a company likes to receive before its market debut.
Le débat sur la sécurité arrive au pire moment pour la valorisation
Timing matters. Anthropic sold 980 million shares at 83 dollars each, bringing the operation size to over 81 billion dollars, according to CNBC. The fully diluted valuation stands around 1.5 trillion dollars [to be verified], a level placing the company at the center of market expectations. Reuters also reports that demand for the offering was about six times the available shares.
At this stage, the case establishes neither legal violation nor proven danger from the company’s models. It mainly scrutinizes the consistency between Anthropic’s public message and its industrial strategy. The conflict is broader than a simple ego duel in Silicon Valley: it concerns how investors, regulators, and customers assess safety promises in cutting-edge AI.
Des chiffres vertigineux, mais aussi des obligations lourdes
The financial figures presented in the prospectus explain why the case attracts so much attention. Anthropic declared 33.24 billion dollars in revenue over the nine months ending September 30, a 1,200% increase year-on-year, according to CNBC. The company also reported about 17 billion dollars in revenue for the third quarter, up 47% from the previous quarter, again according to cited documents.
But the other side of the story is less light. The prospectus states about 900.9 billion dollars in purchase and lease commitments, mostly related to computing and infrastructure. Approximately 451.6 billion are due before the end of 2030, and some 70.7 billion are due within 12 months following September 30, 2026. These amounts show one simple thing: at Anthropic, the bill rises almost as fast as the ambitions.
- 89 signatories at the statement launch, then 92 on November 19.
- 83 dollars per share, IPO price set.
- 980 million shares sold, according to CNBC.
- 33.24 billion dollars in revenue over nine months, as of September 30.
- 900.9 billion dollars in infrastructure commitments declared in the prospectus.
Pourquoi cette attaque pèse sur tout le marché de l’IA
The Anthropic case interests far beyond its shareholders. An IPO of this size can serve as a valuation benchmark for other frontier AI players, those most advanced and costly models to develop. It can also influence cloud providers, chip manufacturers, and enterprise clients who allocate budgets among multiple tools.
The signatories of the "MisAnthropic Statement" specifically challenge this posture of influence. According to them, the company seeks to accumulate power in the name of the common good, while leveraging its reputation for caution to raise funds and sway the rules of the game. For their part, Anthropic and its founders defend an opposite mission: to help humanity "chart a safe course through this technological evolution," according to the prospectus. Between the two, the market will have to decide on a very down-to-earth point: who pays, for what, and under what conditions.