Anthropic prepares wider rollout of Mythos as it moves toward an IPO

In short: Anthropic is growing fast and planning an IPO while expanding access to Mythos, a powerful AI model it says needs careful control.

What's going on

Anthropic, the company behind the Claude chatbot, is trying to balance its public promise of building safer AI with strong pressure to grow. The Financial Times reports this tension has become more visible as the company works more closely with governments and large businesses.

One flashpoint is a dispute with the US Department of Defense. Anthropic says the Pentagon asked for its AI system to be available for “all lawful purposes,” and the company asked for two exceptions: mass surveillance of Americans and lethal autonomous weapons. After public attacks from President Donald Trump and Defense Secretary Pete Hegseth, Anthropic sued the Pentagon in March, and the case is still in federal court.

At the same time, Anthropic is pushing ahead with a new AI model called Mythos. Analysts quoted in the report describe it as a very strong hacking tool, meaning it can find weak spots in software (like a burglar spotting unlocked windows). Anthropic says its own internal testers found Mythos could identify and exploit vulnerabilities across major operating systems and web browsers. Instead of releasing it widely to the public, Anthropic plans to expand access to 150 organizations in 15 countries that run “critical infrastructure,” such as banking, power, water, and healthcare.

The company’s business is also accelerating. The report says Anthropic announced a $65bn funding round and filed paperwork for an initial public offering, which could value the company at around $1tn.

What to watch

Investors and regulators will be watching whether Anthropic can keep its safety limits in place as it grows, especially if it becomes a public company and faces more pressure to move quickly.

Source: Financial Times