New York City Comptroller Urges Anthropic to Delay IPO, Citing Unresolved AI Risks
New York City Comptroller Mark Levine has publicly called on artificial intelligence company Anthropic to delay its initial public offering in the United States, arguing that the AI startup has not adequately addressed "unprecedented risks" across system safety, corporate governance, and the pace of technological development. The demand adds external pressure on Anthropic as it prepares for a potentially blockbuster listing, while also thrusting risk-management concerns surrounding advanced AI systems into the capital markets spotlight.
According to a letter obtained by media outlets, Levine explicitly stated in his correspondence to Anthropic CEO Dario Amodei that the company should first address these risks before proceeding with its listing. "Rather than rushing headlong into an IPO, the company should focus its efforts on confronting these unprecedented risks," he wrote.
Neither Levine's office nor Anthropic responded to requests for comment as of press time.
Levine's official role makes this appeal particularly noteworthy. The New York City Comptroller oversees municipal finances, manages public spending, and participates in the operation of the New York City public pension funds. Market sources indicate that the city's public pension funds currently hold approximately $7 million (roughly NT$230 million) in equity exposure to Anthropic. This means Anthropic's governance quality and risk controls are not merely an internal matter for a Silicon Valley startup—they directly implicate the long-term fiduciary responsibilities tied to public funds.
Key Concerns: Safety, Governance, and Regulatory Gaps
Levine's primary concerns center on three areas: the safety of AI systems, corporate governance structures, and whether the pace of technological advancement has already outstripped the ability of regulators and risk-management frameworks to keep up.
Over the past year, AI system capabilities have advanced rapidly, drawing increased scrutiny of potential risks including cybersecurity threats and the possibility of technology misuse. Notably, Anthropic's own research has surfaced issues that increasingly autonomous AI models may pose. In controlled testing environments, these systems have demonstrated certain potentially harmful capabilities, including sabotaging code, facilitating fraud, and manipulating information.
Such research findings align with Levine's concerns and raise further questions about how Anthropic balances commercial expansion against its safety commitments.
Massive Listing Scale Alongside Strong Business Momentum
Anthropic's IPO is reportedly on track to proceed as early as November, targeting a valuation exceeding $2 trillion (approximately NT$64.8 trillion). A listing at that valuation would rank among the largest public-market offerings ever. Its primary competitor, ChatGPT developer OpenAI, is expected to go public in 2027.
As it prepares for the listing, Anthropic continues to showcase robust business growth to potential investors. Market sources indicate the company's third-quarter revenue grew 45% from the prior quarter, reaching $16.8 billion (approximately NT$540 billion).
The simultaneous emergence of strong revenue momentum and external concerns over safety governance makes Anthropic's ability to balance commercial expansion with risk control one of the most closely watched issues ahead of its IPO.
The Intersection of Public Funds and AI Governance
Levine's intervention highlights an emerging trend: as AI companies move toward public markets, long-term institutional investors such as pension funds are no longer evaluating only potential financial returns. They are increasingly incorporating non-financial factors—including technology safety, governance arrangements, and regulatory compliance—into their investment decision frameworks.
For funds like the New York City public pension system, which carry long-term payout obligations, holding equity in high-valuation, high-growth, yet high-uncertainty AI startups presents a distinct risk profile. Levine's call for a delayed IPO, made in his capacity as Comptroller, represents a formal statement from a public-fund steward regarding these risks.
As of the time of reporting, Anthropic has not publicly responded to Levine's request, and there is no clear indication as to whether the appeal will affect the company's listing timeline.