NYC Comptroller Urges Anthropic to Pause IPO Amid AI Safety Concerns
New York City Comptroller Mark Levine has formally requested that Anthropic delay its initial public offering (IPO) until the company adequately addresses safety and security risks associated with its artificial intelligence technology. Levine, who oversees pension funds managing $330 billion in assets for over 750,000 city workers, argues that prudence requires pausing the listing process to allow investors to properly assess these risks. Although the city’s current direct exposure to Anthropic is minimal at approximately $7 million, Levine warns that an IPO would force the funds to become long-term investors automatically through index tracking, such as the Russell 3000. This request comes as Anthropic CEO Dario Amodei has previously called for an industry-wide slowdown in developing advanced models, and as prosecutors from eight US states recently sued rival OpenAI for alleged concealment of AI risks.
Key points
- Mark Levine, the supervisor of New York City’s pension funds, sent a letter to Anthropic CEO Dario Amodei urging a pause in the IPO process.
- The request is driven by concerns over the safety and security of Anthropic’s AI technology, which Levine argues must be fully addressed before public listing.
- New York City’s pension funds manage $330 billion in assets and provide retirement benefits to more than 750,000 individuals, including teachers, police, and firefighters.
- While current direct exposure to Anthropic is only about $7 million, an IPO would lead to automatic investment through index funds like the Russell 3000, increasing the city’s long-term stake.
- Anthropic CEO Dario Amodei has previously advocated for a slowdown in the development of advanced frontier AI models, aligning with the safety concerns raised by Levine.
- The request follows a lawsuit filed by prosecutors from eight US states against OpenAI, a rival AI company, for allegedly concealing risks and implementing inadequate safeguards.
Background
Anthropic had previously postponed its IPO to November, citing existential risks and high operational costs in its prospectus. In September, both Anthropic and OpenAI CEOs called for a coordinated slowdown in frontier AI development, with OpenAI confirming it would not pursue an IPO in 2026. Earlier in October, Anthropic’s IPO filing revealed a $42 billion financing arrangement with Broadcom to fund tensor processing unit leases, highlighting the massive capital requirements of the AI sector. The current request from New York City adds a new layer of regulatory and financial scrutiny to Anthropic’s public listing plans, reflecting growing concerns among major institutional investors about the safety and stability of advanced AI technologies.
Why it matters
The intervention by a major US pension fund signals that institutional investors are increasingly prioritizing AI safety and risk management over rapid market entry. This could set a precedent for other large investors to demand similar delays or safeguards from AI companies seeking public listings, potentially slowing the pace of capital inflow into the sector. It also highlights the tension between the aggressive expansion of AI companies and the growing demand for regulatory oversight and risk mitigation, which could influence future policy and investment strategies in the technology sector.
What to watch
Anthropic is expected to respond to Mark Levine’s letter, potentially adjusting its IPO timeline or providing additional assurances on AI safety. The outcome of this request could influence the decisions of other major investors and regulatory bodies regarding the public listing of AI companies. Additionally, the ongoing lawsuit against OpenAI may set a legal precedent for how AI companies are held accountable for risk disclosure and safety measures, further impacting the landscape for upcoming IPOs in the sector.
Want the full story? Read the original reporting
Read on France 24