New York City’s comptroller wants an Anthropic IPO pause until the company shows how it will handle the AI risks it keeps warning about. Mark Levine, who oversees more than $325 billion in city pension money, sent the request straight to CEO Dario Amodei. Now the two sides are talking, just weeks before what could be one of the largest stock debuts ever.
Why the Comptroller Wants an Anthropic IPO Pause
Levine’s letter makes a simple argument. Anthropic has spent recent weeks telling the public that frontier AI may be moving faster than society can understand or control it. If that is true, he says, the company should fix its safety and governance plans before it takes on the pressure of being a public company.
He points to Amodei’s September 12 essay, “We Must Pace the Frontier,” where the CEO said the industry needs to slow how quickly AI models gain new abilities. Levine argues that the same caution should apply to the timing of the IPO. Getting ready for a listing, he writes, pulls money and attention away from safety work.
What the Letter Asks For
The comptroller is not just asking for a delay. He wants written answers to a few specific questions:
- How Anthropic will set up audits and controls for its AI systems
- What risks those controls are meant to cover
- A projected timeline for putting them in place
- Whether those controls actually support Amodei’s own call to slow AI progress
He also called the voluntary safety accord that Anthropic and five other companies signed with the White House too weak for the moment. In his view, a voluntary pledge falls short of the governance the risks call for.
Anthropic IPO Talks Begin as the Listing Nears
Levine confirmed on Bloomberg Television that Anthropic had reached out and that discussions had started, though he would not share details. He said the company seemed to be under pressure from several directions. His main worry is simple: “I worry about going public and then having an incident,” he said.
The timing is tight. Anthropic told prospective investors this week that third-quarter revenue hit $16.8 billion, up 45% from the previous quarter. The company is reportedly preparing for a listing as soon as November.
| Detail | What We Know |
|---|---|
| Who is asking | NYC Comptroller Mark Levine |
| Pension assets he oversees | More than $325 billion |
| Anthropic Q3 revenue | $16.8 billion, up 45% from Q2 |
| Reported valuation talk | About $1.5 trillion to $2 trillion |
| Possible listing | As soon as November (reported) |
| Status | Talks have begun, no details shared |
You can read more about those numbers in our look at Anthropic’s record quarter. Some investors have reportedly floated a valuation closer to $1.5 trillion, below earlier talk of about $2 trillion.
The Risk Warnings Behind the Request
Levine leaned heavily on Anthropic’s own words. He cited reports that the company’s draft S-1 filing warns of catastrophic or even existential risks from advanced AI, with risk sections running about 80 pages. He also noted plans for a governance setup that may give founders and trustees concentrated control.
He referenced the September resignation of an Anthropic researcher who warned about AI’s dangers, too. Together, he argues, these signals describe risks that no normal IPO has ever had to disclose.
Amodei Wants Government Power to Block Unsafe AI
While the IPO debate plays out, Amodei is pushing hard for tougher rules. In a new interview with ABC News, he said Anthropic is proposing that the government get the power, in a narrow way, to block the release of unsafe AI. That follows a letter he released this week saying a model could be blocked if third-party testing finds it poses unacceptable risks.
Alongside that letter, Amodei announced a $200 million investment in research on how AI affects society. He also said the job market may need help this time, from better unemployment insurance to incentives that keep companies from laying people off. For the longer term, he floated ideas like higher capital gains taxes or taxes on AI companies.
On the idea of pausing AI development, Amodei said a pause would only work if every major player agreed and it could be verified. He added that he does not trust China to take part responsibly. Anthropic’s co-founder Jack Clark made a similar case this week in his call for regulation.
What This Means for You
If you use Claude every day, nothing changes for you right now. But this fight matters because it tests whether AI safety promises hold up once Wall Street money arrives. Here is what to watch:
- Whether Anthropic gives Levine a written answer and shares it publicly
- Any change to the reported November listing plans
- Whether other big pension funds join the call for a delay
- How Washington responds to Amodei’s push for power to block unsafe models
Regulators are already paying attention, as shown by the recent FTC safety probe. Anthropic has also opened its systems to outside safety auditors, which may be part of its answer.
Our take: Anthropic built its brand on warning about AI risk. Now an investor is asking it to apply those warnings to its own IPO, and the answer will say a lot about how much those warnings really guide the company.
For more daily coverage, visit our AI news hub.
Sources
This report draws on the NYC Comptroller’s letter to Anthropic, coverage from Crypto Briefing, and an interview with Dario Amodei by ABC News.
Conclusion
New York City’s comptroller is asking Anthropic to put its IPO on hold until it explains how it will control the risks it has publicly flagged. Anthropic has opened talks, but neither side has shared what was said. At the same time, Amodei is pressing for government power to block unsafe AI and pledging $200 million to study its effects on society. With a listing reportedly possible in November, the company now has to show investors and the public that its safety talk and its business plans point in the same direction. How it answers could shape how every AI lab approaches going public.
