New York City Comptroller Mark Levine submitted a formal letter to Anthropic asking the artificial intelligence company to pause its planned public offering until it had resolved open questions about safety practices, corporate structure, and the pace of its technology deployment, according to a news outlet. On Wednesday, Levine directed a written appeal to Anthropic's chief executive, Dario Amodei, outlining the request.
Overseeing the city's finances falls within that office's mandate, which also extends to stewarding public pension portfolios whose combined assets top $325 billion. That Thursday, Levine appeared on a prominent broadcast platform to respond to the letter and speak to his efforts to postpone the offering.
In a written assessment, Levine argued that the pledge on AI safety negotiated between the Trump administration and a group of AI companies — Anthropic included — fell short of what the situation required. Levine urged Anthropic to establish a governance structure proportionate to the risks that the company and others — including researchers working in the field — had themselves identified, calling in a written statement for the organisation to detail with precision how it plans to carry out audits and controls, what those measures are designed to cover, and when they are expected to be in place, while also explaining whether the controls would adequately respond to the company's own position on moderating the rate at which AI model capabilities are developed.
No reply from Anthropic had come through by the time the outlet published its story.
On Wednesday, Anthropic disclosed to potential backers that its revenue for the third quarter had come in at $16.8 billion, with the company also posting a profit on an adjusted basis. Plans call for an IPO roadshow to get underway around the week of Nov. 9, with the public offering itself expected to land before year-end — though no formal timeline has been locked in. The planned share sale targets a scale matching or surpassing what SpaceX raised through its flotation that June.
Amodei, whose remarks Levine drew on, put forward a call last month for AI firms to pull back on the pace at which their most advanced models are being developed, while also pushing for federal regulators to take a more active role in overseeing the industry. A gathering of technology company leaders, including representatives from Anthropic, took place at the White House on Sept. 29 to discuss artificial intelligence safety with the president. Out of that gathering came a pledge — voluntary in nature and carrying no binding force — under which each participating company bears individual responsibility for ensuring its own technology is developed safely.
