Going public puts Anthropic’s safety mission under new pressure
After months of speculation and even a blacklisting from the U.S. military, Anthropic has announced it’s going public. The company said it has filed paperwork with the Securities and Exchange Commission for a proposed initial public offering of its common stock.
The company made the announcement in a brief blog post on Monday, writing that the SEC is reviewing its offer. The post did not provide details on the number of shares Anthropic is offering.
If approved, the company, already valued at nearly $1 trillion, would no longer exclusively rely on private capital. Anthropic would have a vastly larger pool of revenue and reduce its reliance on single tech megacorporations like Google and Amazon.
What does a public AI company look like?
A public Anthropic would look very different from what it does now, and that’s because it’s currently a public benefit corporation. That means it’s explicitly designed to prevent short-term profit motives from affecting safety priorities.
An IPO would likely compel the company to choose between prioritizing shareholder interests over its current mission and implementing a dual-class share structure. This latter option, similar to Google’s model, would allow founders to maintain voting authority even with a smaller financial stake.
Anthropic would also look different on paper compared to other, more traditional tech companies. The company would likely show lots of revenue, but it would also likely show that it’s still burning massive amounts of cash on AI compute.
AI companies have seen astronomical valuations not because of profit, but because of the technology’s possibilities. That means margins can potentially be very high, but they are nowhere near that now.
The amount of revenue used on research and development is also extreme compared to other tech giants. Google, for instance, spends up to 15% of its revenue on R&D, but Anthropic reports using almost all of its revenue on it. Company leaders predict that research costs will drop to 65% of revenue in the coming years.
Internal conflict will also rise if the company goes public. If its biggest competitor, OpenAI, releases a more powerful chatbot, analysts could consider Anthropic slow and too safe, even if the company believes it made the right call. This could result in some brutal quarterly earnings calls with aggressive investors upset that the other teams beat them.
Does Anthropic’s relationship with Trump hurt its IPO chances?
But Anthropic’s public offering isn’t a done deal yet, since the SEC still needs to approve the funding, and previous issues between the company and the government may come back to bite.
In late February, the Pentagon labeled Anthropic a supply chain risk, effectively blacklisting the company. Defense Secretary Pete Hegseth said it was because Anthropic CEO Dario Amodei refused to remove safeguards from the company’s AI model. Hegseth said this effectively allows a private company to dictate how the government uses data.
The defense secretary called the company’s refusal an attempt to “seize veto power over the operational decisions of the United States military.”
Anthropic later sued, alleging that the military was punishing it for failing to comply with its demands. But a judge denied its appeal, and the blacklisting was allowed to go through.
Despite the rocky relationship between the White House and Anthropic, the SEC would likely not deny it. Being banned from working with the U.S. military may push away some investors but entice others.
How does this change the AI industry?
Anthropic isn’t the only AI company trying to go public — OpenAI is, too. But OpenAI’s business structure makes that difficult. The ChatGPT-powered tech giant is currently a capped-profit LLC controlled by a nonprofit, but if Anthropic went public first, OpenAI would almost certainly accelerate towards restructuring into a for-profit business. Since Anthropic depicts itself as the responsible AI company, its IPO gives political cover for other AI companies seeking to follow suit.
Currently, the industry believes that the safe approach to AI is the wrong approach. Move too slowly, and the other side might beat you. But Anthropic has always disagreed with that.
If the market rewards Anthropic’s safety approach with a premium valuation, it indicates that the market prefers safety over speed and doesn’t see it as a hindrance. The broader AI industry is anticipating a definitive market signal regarding the economic impact of safety measures — specifically, whether they enhance or diminish a company’s value. A potential Anthropic IPO is expected to provide that much-needed clarity.
Round out your reading
- Plea bargains keep America’s courts running. Guilt or innocence barely matters.
- Trump says the media isn’t covering one of his biggest accomplishments. We checked.
- Mike Lindell denied MyPillow was hacked. Its private data is now online.
- Red meat allergy rises with ticks as HHS targets Lyme disease, alpha-gal syndrome.
- We’re building a new Straight Arrow. Help us shape our future by taking our survey.
Ella Rae Greene, Editor In Chief
