New York City's chief fiscal watchdog is telling one of the world's most valuable startups to slow down before it goes public. Comptroller Mark Levine sent a letter to Anthropic CEO Dario Amodei this month demanding the company postpone its planned November initial public offering until it resolves pressing concerns about AI safety, governance, and the pace of its own development.
A Letter, a Broadcast Interview, and a Blunt Message
Levine's push went beyond a formal letter. He also sat for a broadcast interview, as reported by CNBC, where he reiterated his call for Anthropic to pause the listing until it can demonstrate more caution and transparency. Finimize also described Levine's letter calling for the IPO to be paused until the company addresses concerns about safety and governance.
Levine is not a random critic shouting from the sidelines. As comptroller, he serves as investment advisor and trustee for New York City's pension funds. That role gives his objections institutional weight whenever a company his funds might invest in prepares to sell shares to the public.
Revenue Soaring, Valuation Climbing Toward $2 Trillion
The financial backdrop makes the standoff especially striking. At an investor presentation in San Francisco on Wednesday, Anthropic reported third-quarter 2026 revenue of $16.8 billion, a 45% sequential jump over the prior quarter, and said it was cash-flow positive, according to Seeking Alpha. That marks a dramatic swing from the $4.6 billion in full-year 2025 revenue and an operating loss of over $8 billion that year. The company has also spent substantially on infrastructure.
Some investors believed a valuation of $1.8 trillion to $2 trillion was possible for Anthropic ahead of the planned November listing, building on a May funding round that already valued the startup at $965 billion, SmartAsset reports. Anthropic filed a confidential draft S-1 registration statement with the Securities and Exchange Commission back in June, setting the stage for the offering now drawing Levine's scrutiny.
The CEO's Own Warnings About Catastrophic Risk
At the heart of Levine's objection is a tension that Anthropic's own leadership has acknowledged publicly. In September 2025, CEO Dario Amodei estimated a 25% chance of catastrophic outcomes from advanced AI. Separately, he has discussed controlled pacing for AI development. The warning followed high-profile internal safety researcher resignations and ongoing debate over the risks posed by increasingly autonomous models.
In fictional tests, models exhibited behaviors including evasion, deception, and attempts to steal corporate secrets.
A Governance Structure Built for Mission, Not Shareholder Control
Anthropic is organized as a Public Benefit Corporation overseen by an independent Long-Term Benefit Trust, a structure intended to let the company legally prioritize AI safety over profit, according to the CLS Blue Sky Blog. The same governance design, however, concentrates director appointment power away from public equity holders — meaning that even as outside investors buy in, they may have limited sway over the board. PBC charters generally permit directors to weigh public mission goals alongside financial returns, but that can constrain the kind of shareholder governance seen at traditional public companies.
Not Levine's First Fight Over IPO Governance
This is not the first time Levine has used his institutional platform to challenge a tech giant's path to public markets. In May, he joined New York State Comptroller Thomas DiNapoli and CalPERS CEO Marcie Frost in a joint letter opposing SpaceX's proposed IPO governance structure, which they argued severely restricted public shareholder voting rights by relying on perpetual super-voting shares and mandatory arbitration rules, per the Office of the New York State Comptroller. The Anthropic letter extends that same pattern of collective investor pressure against mega-cap tech governance terms.
Levine was elected comptroller in November 2025 and took office as the city's 52nd comptroller on January 1, succeeding former Comptroller Brad Lander, after previously serving as Manhattan Borough President and a member of the City Council.
Why This Matters Locally
Levine's interest in Anthropic isn't limited to pension governance. His office's own May report on AI's local fallout estimated that long-term AI-driven workforce displacement could put up to 259,000 New York City jobs at risk and create as much as a $14 billion municipal revenue shortfall, prompting a recommendation to expand the city's rainy-day reserves. That broader concern about AI's economic disruption appears to be feeding directly into how closely his office is now watching the governance and safety posture of companies like Anthropic before they tap public capital markets.
