Anthropic Pushes IPO Plans While AI Safety Rules Advance

Anthropic is preparing an initial public offering that projects substantial future revenue even as its chief executive and other industry leaders urge greater caution around advanced AI systems. California and federal officials are simultaneously considering new oversight measures, including requirements for emergency shutdown mechanisms on powerful models.

The company’s valuation discussions and revenue forecasts come at a time when major AI developers report annual infrastructure spending in the hundreds of billions of dollars. Revenue growth has not yet matched those outlays, according to statements from critics including former BlackRock manager Ed Dowd and commentator Ed Zitron. Zitron noted that OpenAI alone lost $21 billion in 2025, a figure larger than losses recorded by earlier unprofitable startups such as Amazon or Uber over comparable periods.

Dario Amodei of Anthropic, along with Sam Altman of OpenAI and Elon Musk, have publicly warned that AI could become difficult to control. Representative Alexandria Ocasio-Cortez acknowledged the need to address safety but questioned whether the warnings serve mainly to secure government support. Akhil Verghese of Krazimo argued that reported “rogue AI” incidents simply reflected inadequate guardrails rather than independent rebellion by the models.

Policymakers in Sacramento and Washington are examining rules that could mandate kill switches or other containment features. No final legislation has passed, and details of any required mechanisms remain under discussion. The tension between rapid commercialization and regulatory safeguards continues without a settled resolution.