Anthropic, the company behind the Claude AI models, is now talking with New York City Comptroller Mark Levine. That follows a letter in which Levine asked the company to delay its planned US initial public offering.
The timing matters. Anthropic is reportedly targeting a listing as early as November 2026, and one of the country’s most visible pension overseers has asked it to tap the brakes.
What Levine asked for, and why
Levine’s letter is dated approximately October 14-15, 2026. In it, he urged Anthropic to halt its IPO until it adequately addresses concerns about AI safety, governance and the speed of technology development.
His argument has a pointed twist. Levine suggested a public offering could distract the company from risk mitigation work that Anthropic’s own research has flagged as important.
The comptroller’s platform gives the letter weight. Levine oversees New York City pension funds with over $325 billion in total assets.
The city’s direct stake in Anthropic is far smaller. NYC pension funds have exposure of approximately $7 million to the company through private equity.
By October 16, 2026, Anthropic and Levine’s office had established communication. Neither side has publicly laid out what the conversations cover or whether any commitments have been made.
The numbers behind the IPO push
Levine’s request lands while Anthropic is posting the kind of financials that make bankers reach for their calendars. During investor meetings, the company reported Q3 2026 revenue of $16.8 billion.
That figure represents 45% growth quarter-over-quarter.
Anthropic also disclosed that it is cash-flow positive.
A potential IPO valuation as high as $2 trillion has been discussed, though that figure is a possibility, not a set price.
The company confidentially filed for its IPO in June 2026. A confidential filing lets a company work through regulatory review privately before going public with its paperwork.
So far, no changes to the IPO timeline have been made public. The November target, as reported, still stands.
Background: safety meets the stock market
Anthropic has built much of its identity around AI safety, which makes Levine’s critique unusually targeted. He is not accusing the company of ignoring risk so much as asking whether going public could pull attention away from its own stated priorities.
The legal entity is Anthropic PBC, a public benefit corporation designation. That label already signals the company is meant to balance shareholder interests with broader goals, a tension an IPO would put under a much brighter spotlight.
It also sits inside a competitive race. OpenAI is among the peers also contemplating public offerings, so the question of who lists first, and on what terms, carries strategic weight.
What this means for Anthropic and its future shareholders
The engagement itself is notable. A company with $16.8 billion in quarterly revenue does not have to answer a letter from an official whose funds hold approximately $7 million of its exposure, yet Anthropic opened a line of communication within days.
Key things to watch include any public statement from Anthropic or Levine’s office about the discussions, any movement in the November target, and whether the public version of the IPO filing includes expanded disclosures on AI risk and governance.
Disclosure: This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our Editorial Policy.