Anthropic CEO Dario Amodei previously publicly called on artificial intelligence industry participants to slow down technology development to reduce the likelihood of unpredictable consequences. However, six weeks after OpenAI and SpaceX executives publicly backed the initiative, the tech race not only failed to stop but continued with renewed vigor. Over this period, Anthropic, OpenAI, Google, and SpaceX unveiled new AI models, while Anthropic itself filed for an IPO that could be the largest in history. In its filing, the company stated its intention to channel hundreds of billions of dollars into developing the Claude chatbot and compute infrastructure, raising questions from AI safety advocates who insist on evaluating companies' actual actions rather than their executives' public statements.
From Slowdown Call to Record Offering
According to disclosures in the offering documents, Anthropic’s business is demonstrating explosive growth: for the first six months of 2026, the company’s revenue reached approximately $16.3 billion compared to about $1.16 billion for the same period a year earlier, while reporting a significant loss for the first half of the year. The stock offering is planned on the Nasdaq exchange under the ticker symbol ANT. The combination of record growth rates, plans to spend hundreds of billions on infrastructure, and recent public warnings about catastrophic AI risks has triggered heightened criticism from regulators and specialized organizations.
Incidents and Senators' Inquiries
Industry concerns intensified following a series of incidents over the past month: AI-powered cyberattacks targeted companies across multiple countries. In addition, OpenAI reported that one of its AI agents attempted to bypass a monitoring system. Against this backdrop, Senator Elizabeth Warren demanded that Dario Amodei publicly explain how his warnings about catastrophic AI risks align with the company's plans to go public and scale development.
Anthropic's Steps and Friction with the Safety Community
Anthropic itself is taking a number of measures to enhance safety: the company temporarily paused certain work following unpredictable model behavior and began inviting independent researchers into its offices to observe development. Nevertheless, Anthropic's relationship with the AI safety research community is gradually deteriorating. Machine Intelligence Research Institute head Malo Bourgon noted that if the company positions itself as an organization bearing heightened responsibility for the consequences of AI development, experts have the right to expect higher standards precisely in this area.
The Bill and Amodei's Response
Additional criticism was sparked by an AI regulation bill drafted by Anthropic lobbyists: specialized organization representatives believe it could limit the effect of existing state-level laws and contradicts the company leadership's public statements on the need to control AI. In response to the criticism, Amodei stated that Anthropic's strategy is a “race to the top” in standards and an intention to show how AI can be developed responsibly, encouraging competitors to follow its example. He also recalled decisions that cost the company dearly, including a conflict with the US Department of Defense over refusing to allow its technologies to be used for domestic surveillance and fully autonomous weapons. Meanwhile, independent tests show that modern chatbots are less likely to support dangerous user misconceptions and more frequently discourage potentially harmful behavior.
