Why the timing is drawing fire
In a Wednesday Bloomberg Surveillance interview, Levine argued that moving to list now clashes with recent safety concerns. "This is cognitive dissonance right now because we've had months of serious safety incidents," he said, noting Anthropic has acknowledged its technology could pose catastrophic risks. His bottom line: "It's a contradiction that the company is going public right now." Or as he put it more bluntly, "The Anthropic IPO just adds to the stakes."
Pensions, passive exposure, and a prior push to delay
Levine helps manage New York's fiscal health and its public retirement funds, which together control more than $325 billion. In October, he sent a letter urging the developer of Claude to put the IPO on pause. He also flagged that about a third of the city pension system sits in index-tracking strategies, which means those portfolios would end up owning Anthropic once it is included in relevant benchmarks.
The bigger AI risk conversation among pensions
Levine said the city is already deeply invested in AI across private equity, private credit, infrastructure, and developing markets, and that discussions are underway with public pension systems nationwide on how to handle AI risk. The debate is not just in New York. This week, the California Public Employees Retirement System said it intends to step up scrutiny of AI risks. Staff told the board at a Monday committee meeting that cyber threats and whether AI models can be trusted are becoming a larger part of their due diligence.
As of June 30, Calpers' exposure to companies tied to AI was estimated at $65 billion, according to Equable Institute.
When tech governance and investor oversight matter, get clear five-minute context each morning with free Market Briefs
What this could mean for your portfolio
If you own broad market index funds, you may get Anthropic exposure whether you seek it or not, which is why governance questions loom larger here than for a niche tech bet. Levine's stance signals more engagement from big pensions on AI risk, from cybersecurity checks to model reliability. That scrutiny can shape disclosure, oversight, and ultimately how AI names price risk going forward.
Because public accountability and market transparency affect many investors, subscribe to our free Market Briefs newsletter for quick updates
