Anthropic’s $2 Trillion IPO Is Still On. And So Is the AI Race

Key Points

1. Bloomberg Opinion argues Anthropic's Nasdaq listing and frontier research will both proceed despite slowdown talk.
2. Dario Amodei's essay urging firms to pace the frontier drew public agreement from Sam Altman and Elon Musk.
3. At stake: 80% gross margins, $65 billion in annualized revenue, and a race against Chinese rivals.

The latest:

Anthropic’s planned Nasdaq listing, which could value the company at $2 trillion, shows signs of staying on track even as its chief executive calls publicly for slower frontier AI development, Bloomberg Opinion columnist Parmy Olson wrote. Nvidia may contribute roughly $10 billion, she reported. Olson argued the industry has too much at stake to genuinely brake the race to superintelligence.

Details:

• The essay: Dario Amodei, Anthropic co-founder and chief executive, wrote that AI companies must pace the frontier and slow the rate at which they improve model capabilities. Olson noted he simultaneously described a balanced rate of advancement and stressed progress will still seem fast, stopping short of any moratorium.
• The endorsements: OpenAI’s Sam Altman and SpaceX’s Elon Musk publicly agreed with Amodei’s post, according to Olson. She argued neither company is likely to slam the brakes, and that the statements are reassuring only until weighed against how little they may ultimately change.
• The numbers: Anthropic’s margins rest on the promise of future capabilities rather than existing cash flows or its $65 billion in annualized revenue, Olson wrote. David Rainville, lead manager of the Sycomore Sustainable Tech fund, which holds about 800 million euros in global tech stocks, cited 80% gross margins as being at risk.
• The commercial case: A pause on more powerful systems would not hurt Anthropic immediately, Olson argued, since many businesses avoid the most expensive frontier models and the company could save billions on training. Longer term, she wrote, Claude risks commoditization against OpenAI’s ChatGPT and Google’s Gemini.
• The China argument: Amodei made clear he does not want American firms to ease up enough for Chinese competitors to pull ahead, and said ceding ground to what he called authoritarian regimes, chiefly the Chinese Communist Party, would be a security threat, according to the column.
• The White House: President Donald Trump publicly ruled out federal coordination of such efforts, ignoring a request from Amodei, and said the only guardrails AI needed were a strong, high-IQ president, Olson reported. Trump wants the US to beat China to superintelligence as the ultimate way to win.
• The concrete step: Amodei’s most specific recommendation is embedding independent evaluators inside AI companies to verify frontier model work is safe, and Anthropic has already committed to this, Olson wrote. She expects firms to focus there rather than meaningfully decelerate research and development.
• The auditor question: Olson questioned Amodei’s example of an independent evaluator, the non-profit METR, arguing it draws from the same tight-knit AI safety world as Anthropic. She noted researcher Ajeya Cotra spent more than nine years at Coefficient Giving, formerly Open Philanthropy, where Amodei was once a technical advisor.
• The risk methods: Anthropic’s leaders believe superintelligent systems are inevitable and that it would be irresponsible to let the least cautious actors arrive first, Olson wrote. She said this drives Amodei to pursue recursive self improvement, where AI systems largely create their own successors.

Background:

Amodei’s own researchers have warned that AI models could kill all humans as they grow more powerful and gain the ability to redesign and improve themselves, according to Olson. The UK has been testing frontier models through its AI Safety Institute since 2023.

Between the lines:

Olson’s reading turns on a single word: pacing a technology can mean controlling its speed toward a goal rather than slowing it, and she cited marathon runner Eliud Kipchoge, who used pacemakers and still broke two hours. Read alongside the $2 trillion valuation, the 80% margins and the warning about Chinese rivals pulling ahead, she argued Amodei wants to buy time, not abandon the destination.

What’s next

Watch the Nasdaq listing timeline, whether Nvidia’s roughly $10 billion investment is confirmed, and whether Anthropic or OpenAI name evaluators outside the AI safety network Olson described, such as the UK institute or the US Center for AI Standards and Innovation.