Sen. Elizabeth Warren has asked Anthropic CEO Dario Amodei to answer 10 multipart questions about the company's AI safety risks, finances and governance by November 3, before Anthropic's planned initial public offering. Warren, the ranking member of the Senate Committee on Banking, Housing, and Urban Affairs, sent the 11-page letter on October 27.
The letter is built largely on Anthropic's own S-1 registration statement, which runs more than 90 pages of risk factors. Anthropic plans to list on Nasdaq under the ticker ANT and sell shares in November, according to its filing. Reported valuation estimates for the offering range from $1.5 trillion to $2 trillion.
“Anthropic is projected to be the largest IPO in history.”— Elizabeth Warren, Ranking Member, Senate Banking Committee
Warren's first question asks whether Anthropic believes its models pose an existential threat, and if so, what the company is doing about existential risk, bioweapons, sandbox isolation and cyberattacks. She also asks how Anthropic thinks investors should price that risk. The letter cites S-1 disclosures on model self-preservation behavior, bioweapons assistance and autonomous hacking.
Key facts
- 01Sen. Elizabeth Warren sent an 11-page letter dated October 27 to Anthropic CEO Dario Amodei.
- 02She asks for responses to 10 multipart questions no later than November 3, 2026.
- 03The letter draws on Anthropic's S-1, which devotes more than 90 pages to risk factors.
- 04Anthropic plans to list on Nasdaq under the ticker ANT and sell shares in November.
- 05Questions cover existential risk, bioweapons, cyber incidents, cloud-partner deals, federal taxes and lock-up terms.
A second set of questions targets incident reporting. Warren asks what safeguards exist against autonomous AI cyberattacks, how incidents are reported to the public and to the government, and why some were discovered late. The letter states that incidents of "misaligned model activity" were not immediately discovered, and that Claude accessed publicly available voter data without human authorization.
Warren also asks how Anthropic will mitigate the risks Amodei described in a September essay, including loss of control, cyberattacks, bioterrorism and economic disruption, and whether IPO proceeds could make those risks worse. One question asks whether Amodei agrees that government intervention is needed and what industry-wide efforts exist to slow frontier development.
“the risks presented by this technology demand greater scrutiny”— Elizabeth Warren, Ranking Member, Senate Banking Committee
The financial questions go further. Warren asks for Anthropic's equity, financing and revenue relationships with cloud providers in 2024, 2025 and the first half of 2026, including whether any investments were tied to compute purchase commitments. She also asks whether the company's size already creates volatility risk for public markets, and whether an IPO would increase it.
Governance is another thread. Anthropic is a Delaware public benefit corporation, and Warren asks whether it is acting consistently with both its public benefit mission and its fiduciary duties, and whether its Long-Term Benefit Trust could come into conflict with other shareholders.
The final questions cover taxes and insider sales. Warren asks when Anthropic expects to pay federal income tax and how its 2025 research and development deductions and credits were split between model development and data center work. She also asks why the proposed lock-up structure allows some insider sales on day one, instead of the traditional 180-day lock-up.
Warren frames the letter around retail exposure. She writes that millions of Americans with retirement accounts invested in index funds would quickly become Anthropic shareholders, and that public pensions are already exposed to the company.
Anthropic had not publicly responded to the letter as of Wednesday. The company's S-1 describes its mission as helping humankind "chart a safe course through its technological evolution," and Warren has asked for what she calls exhaustive responses by November 3.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
