Hardware Ready Ready to Bench?
Anthropic IPO. New York Comptroller Mark Levine has asked Anthropic to postpone the planned initial public offering (IPO) until the company provides independent security audits, a more robust governance structure, and a credible plan for managing the development of advanced artificial intelligence models.
The request is contained in a letter sent to Anthropic CEO Dario Amodei. Levine, the financial officer and auditor of the City of New York, highlights that pension funds managed by his administration currently hold a small stake in Anthropic, worth approximately $7 million.
Anthropic IPO: why it matters
Levine explained that if the company were listed on the stock exchange and included in equity indices such as the Russell 3000, funds replicating these indices would automatically purchase its shares. For this reason, he seeks greater guarantees before the listing.
The Comptroller refers to the article “We Must Pace the Frontier,” published on September 12, in which Amodei himself advocated for slowing down the improvement of AI model capabilities. Levine counters that if the pace of development exceeds the capacity for understanding and control, the listing timeline should also be reconsidered.
In the letter, published by the New York Comptroller’s office, Levine asks Anthropic to specify how it intends to implement audits and controls, what risks they must cover, and on what timelines. He also requests clarification on whether such measures are sufficient to put Amodei’s call for a reduction in the pace of AI capability advancement into practice.
The concerns raised are not theoretical, as research conducted by Anthropic itself has highlighted potential harmful capabilities of increasingly autonomous AI models, such as code sabotage, facilitation of financial fraud, and information manipulation. The Comptroller believes these risks warrant verifiable responses before the Wall Street debut.
What changes and what are the effects
The request comes at a time when Anthropic’s IPO is at the center of financial debate, with some market sources indicating a possible listing window in November. However, the company’s valuation is no longer considered certain: some investors estimate it around $1.5 trillion, lower than the approximately $2 trillion previously anticipated, also due to a slowdown in revenue growth.
Levine does not limit himself to technical security but also questions the adequacy of future governance towards public shareholders and asks Amodei for a written response, as well as direct dialogue. The message is clear: before the IPO, Anthropic should demonstrate that its safety commitments have scope, timeline, and accountability mechanisms.
The topic fits into the broader debate on the costs and material limits of AI: Anthropic has repeatedly indicated data centers, energy, and qualified personnel as bottlenecks for the growth of its systems, an aspect that makes the balance between commercial expansion and risk management even more relevant.
Source and further reading on Anthropic IPO: original article.
* Content created with the assistance of artificial intelligence systems.