Anthropic Should Delay IPO to Address AI Safety Risks, Says NYC Comptroller

Author: Bloomberg Television · Watch on YouTube

Video authored by the channel shown, not by IANIA; all rights belong to the creator. Transcript and summary generated automatically; the original-language content is preserved.

Original language: English

Original video: Anthropic Should Delay IPO to Address AI Safety Risks, Says NYC Comptroller · Channel: Bloomberg Television · 06:30 · Watch on YouTube

Topics: AI governance and regulation, Markets and economy, AI models, AI safety

Summary

New York City Comptroller Mark Levine is urging Anthropic to delay its planned IPO, arguing that the listing process and the pressures of public markets could distract from AI safety work. He says New York City pension funds and other public investors have a significant stake and can use their influence, though they cannot legally block the IPO. Levine wants clearer safety thresholds, governance mechanisms, and transparency before the company enters public markets. He supports frontier AI companies eventually going public, but says it should happen only when safety concerns are properly addressed.

Key points

Levine says Anthropic should focus on addressing AI safety risks before proceeding with an IPO.
He argues that preparing for an IPO would consume leadership attention and that public-market growth and profit pressures could compete with safety work.
New York City pension funds could become major Anthropic shareholders through passive index investments if the company goes public.
Levine says investors need more clarity on safety tripwires, capability thresholds for risks such as bio and cyber, and governance that could allow someone to halt development.
The comptroller acknowledges that his funds have no direct legal or regulatory authority to stop an IPO, but says they can exert influence alongside other public pension investors.
He worries a serious safety incident after an IPO could damage Anthropic’s stock and reverberate through markets.
Levine supports frontier AI companies eventually going public, citing transparency and broader public participation, provided it is done safely.
He criticizes index providers for allowing companies to enter indexes quickly, saying this leaves little time to assess concerns.
Levine says Anthropic and OpenAI set the pace for frontier AI and calls for transparent third-party testing.
He takes some relief from OpenAI having paused its IPO, but says he does not know the reason for the pause.

Chapters

Start	Chapter	Summary
00:00	Why Levine is calling for an IPO delay	Levine says Anthropic should address AI safety risks before an IPO, which could distract leadership and create pressure to prioritize growth and profits.
01:39	Investor leverage and unanswered safety questions	He describes the limits of shareholder influence and calls for clearer safety thresholds, timelines, and governance before investors commit.
03:21	Influence, market risks, and eventual public listing	Levine says public pension funds can exert influence and warns that a post-IPO safety incident could harm both Anthropic and the wider market.
04:25	Index inclusion and rapid listings	He supports a safe public listing but criticizes index providers for dropping earlier safeguards and potentially adding companies within days.
05:27	Frontier AI standards and OpenAI’s pause	Levine says Anthropic and OpenAI shape industry standards, advocates transparent third-party testing, and hopes Anthropic follows OpenAI’s IPO pause.

Quotes

An IPO is about the most all consuming thing that corporate leadership could go through. — 00:35

There’s no timeline. — 02:14

Much better to take a little more time right now and get this done. — 03:54

I am not opposed to public listing of these companies, but it’s gotta be done right. — 04:25

I hope Anthropic follows soon. — 05:58

Concepts and entities

Anthropic, Mark Levine, Dario Amodei, OpenAI, SpaceX, Russell 3000

Full transcript (original language)

00:00 Let’s get to the most read letter on Wall Street in the last twenty four hours. The New York City controller Mark Levine are calling for Anthropic to pause its impending IPO over AI safety concerns, writing in a letter addressed to Anthropic CEO, Dario Amin, quote, instead of proceeding headlong with its IPO, the company should focus its energies on addressing these unprecedented risks. Mark Levine joins us now for more. Mark, good morning. Good to see you. Jonathan, Annmarie, thanks for having me. Thanks for making time for us. What prompted the letter? Let’s start there. What changed for you? This is a company building a technology that poses a potential existential threat to humanity. Those

00:35 are not my words. Those are the words reportedly of the s one that they have published. They also do not claim to have solved that problem, posing risk for investors. And and I am an investor, a shareholder, and will be a bigger one if this goes public. We’ll talk about why that is. An IPO is about the most all consuming thing that corporate leadership could go through. It it it will dominate their lives. It will distract from the safety work. And then once they are public, there’ll be new pressures, pressures for growth, pressures for profits that will continue to push against the safety work. And we do

01:08 have skin in the game in the New York City pension funds. We have a little bit of… In the private market right now. But because we assume the indexes are going to rush them in, SpaceX opened that door. We are huge passive investors. We have, like, a 120,000,000,000 in passive investments. We track the Russell three thousand. We are then going to become potentially within days major owners of this company. So we have a stake here, and we want them to take their time, deal with the safety work, mitigate the risk, let investors know we can be confident, and then proceed.

01:39 They know that too. So how much leverage do you have when they know that ultimately you’ll be biased if they go public? Look. I think we have probably more leverage now. It’s why I’m speaking out. Once they go public, we will be shareholders. It’s true. Shareholder powers have been diluted in this era, unfortunately. I don’t think we know all the details on the governance structure of Anthropic. For SpaceX, shareholders were totally disempowered. We were… We are still very upset about that. This is, an accelerating trend, particularly in tech companies. But we’re still gonna advocate at that point. But

02:14 I’m speaking out now because I think there’s still time to address these issues. Yes. We have the White House, accord a step forward. A few years ago, we probably would have thought that was a big deal, but it is nonbinding. There’s no timeline. So what? They have ten years to implement it? And it leaves a lot out, which I’m hearing from other investors. Like, what are the tripwires on safety? What are the capability thresholds you’re worried about with bio and cyber, etcetera? Is there any governance structure that would allow someone to hit the stop button even if there is pressure from competitors? I don’t have those answers.

02:48 Investors don’t have that answer… Those answers. You want that kind of transparency before you hit the public markets. Has the company responded? We did hear from them yesterday. I can’t say there’s a lot of substance yet, but we have a line of communication open. Well, what did they… What was their initial reaction? What did they say? I I… I’m looking… I can characterize the discussions yet, but they they are Do do you find that it was hopeful to your cause? I don’t know yet, but it does appear they’re under pressure from multiple sides right now. So you don’t actually have direct legal or regulatory authority to stop an IPO?

03:21 We do not, but we are one of the largest public investment funds in America and the world. We’re close to a third of $1,000,000,000,000. We own everything in the global economy. And more than that, we’re teaming up with other like minded public pension investors. And so I think we do have influence, and I hope we’re heard. Do you think that influence is going to mean Do you genuinely think they’re gonna postpone it? I don’t know. And I’ll tell you another reason why I really hope so. If they go public and then there’s a catastrophic safety incident beyond… Far beyond what we’ve seen right

03:54 now, That would be terrible for their stock, and I worry that it could then reverberate throughout the markets. Some people are making the opposite point that if they delay their IPO, then it could… That itself could be a problem for the market. And if there’s frothiness, that could be the the trigger. I worry about going public and then having an incident. Much better to take a little more time right now and get this done. By the way, I support frontier AI companies ultimately going public because I want democratic control of world

04:25 changing technology. I want the transparency. I want the broader public to have a piece of this. So I am not opposed to public listing of these companies, but it’s gotta be done right. It’s gotta be done when it’s safe. Have you spoken to the index providers? We have been pushing back on them so hard. What they did with SpaceX now is set a terrible precedent. Remember we used to have a period of seasoning? Let’s watch volatility. Remember there used to be thresholds on the float? Let’s see what insiders are doing. Are

04:55 they dumping all their shares? That’s thrown out the window now, and it could potentially be days, business days, like five business days before this is listed. No time to track any of these concerns, safety being one of them. And, again, because we and so many investors are new… Now doing passive strategies track to indexes. In our case, I say… I think I mentioned we tracked to the Russell three thousand, so that’s… London Stock Exchange owns that. We’ve been pushing very hard for… Against them and many other institutional investors. They can’t resist

05:27 the money, I guess, or the pressure. I think Elon Musk might even been political pressure, but now we’re stuck with this. This is gonna keep happening. Do you feel like it’s a race to the bottom? There are… Look. There there are… We have a duopoly at the frontier right now. Okay? Everyone else is behind. It… This is about OpenAI and Anthropic primarily, and I think that they are setting the pace and the culture and the standards. And if they get this right, if they really implement third party testing with some transparency so we know what the

05:58 testers find, etcetera, that it will, in a very positive way, spread at least throughout the American companies. China is a tougher problem, but they’re still behind. This is super complicated, but I think that one of the two members of the duopoly going public now would make it even more complicated. It will add more pressure. And as you know, the other member of that duopoly, OpenAI, has hit pause on their IPO. They did not say why. They did not attribute this this to safety. I don’t know if it was financial or other

06:30 concerns. I took a little bit of relief from that, especially after the summer’s safety incidents. I hope Anthropic follows soon. Are you also in touch with OpenAI and their leadership? We haven’t made a formal, submission to them, so, I can’t say in any formal way. But in… Informally, talked to the government affairs people of all these firms. Mark, keep us in the loop, and let’s know what the company says. Thank for being here.