TL;DR
- Bloomberg’s The Circuit with Emily Chang went inside Anthropic in a rare, in-depth episode released June 10, 2026.
- Dario and Daniela Amodei discuss the founding story, the Pentagon dispute, and why they say safety and commercial success are the same bet.
- Anthropic is now valued at $965 billion, eclipsing OpenAI’s $852 billion for the first time, after an 80-fold revenue surge in Q1 2026.
- The Pentagon story is not PR - Anthropic refused to remove safety guardrails from its military contract, was blacklisted by the Trump administration, and sued. A federal judge sided with Anthropic.
- A confidential S-1 IPO filing in June 2026 means this stops being a private company conversation soon.
The Bloomberg Documentary: Emily Chang Inside Anthropic
Bloomberg’s The Circuit has done this kind of access piece before - Zuckerberg, Musk, Jensen Huang. But the Anthropic episode feels different in tone. Emily Chang is not sitting across from a founder who has already won. She is sitting across from two founders in the middle of one of the most consequential moments in the company’s short history: record valuation, Pentagon litigation, IPO on the horizon, and model releases arriving fast enough that the competitive landscape changes every few months.
Dario and Daniela Amodei rarely do this kind of access-driven television. The episode earns its runtime.
How It Started: The OpenAI Split
The founding story of Anthropic is worth revisiting because it shapes everything that comes after. Dario Amodei was VP of Research at OpenAI. Daniela was VP of Safety and Policy. In 2021, they left - along with several colleagues - to start a competing lab. The stated reason: a conviction that AI scaling was moving faster than safety research, and that this gap needed a company organised entirely around closing it.
That framing - that safety is not a constraint on progress but the thing that makes progress worthwhile - has not changed in the five years since. What has changed is the company’s commercial position. When Anthropic started, “safety-first” was a differentiator that read as either principled or naive depending on who you asked. In 2026, it is a billion-dollar commercial bet that appears to be paying off.
The public benefit corporation structure they chose is relevant here. Anthropic is not obligated to maximise shareholder return above all else. The board can legally prioritise the mission. That legal structure is easy to dismiss as corporate window dressing. The Pentagon fight is the reason it probably is not.
The Pentagon Stand-Off
This is the part of the Bloomberg episode that matters most. Not because the specific legal dispute is the most important thing Anthropic has done, but because it is the cleanest test of whether the safety rhetoric translates to actual behaviour under real pressure.
The sequence of events: In July 2025, the Department of Defense awarded Anthropic a roughly $200 million contract, making Claude the first frontier AI system cleared for classified military use. Anthropic agreed on the condition that its models would not be used for fully autonomous weapons systems or mass surveillance. That condition held for several months.
Starting in January 2026, the Pentagon pushed for unrestricted use. Anthropic refused. In February, the DoD used Claude in its Venezuela intervention and Anthropic threatened to walk away from the contract. The Trump administration responded by designating Anthropic as a national security supply chain risk - a designation that would effectively bar it from federal contracting entirely.
Anthropic sued. In March, Federal Judge Rita F. Lin granted a preliminary injunction against the government, writing that the DoD had designated Anthropic as a supply chain risk because of its “hostile manner through the press” - and that punishing a company for public scrutiny of a government position is “classic illegal First Amendment retaliation.” The appeals court upheld the designation for covered systems in April, so the litigation continues. The Pentagon has since signed AI contracts with Google, OpenAI, and xAI - companies that did not insist on the same guardrails.
That is the actual story. A company left $200 million and classified access on the table rather than remove constraints on autonomous weapons use. Secretary Hegseth called it “woke AI.” A federal judge called the government’s retaliation unconstitutional. Reasonable people can disagree about the right policy. What is harder to dispute is that Anthropic did what it said it would do.
Dario’s framing in the Bloomberg episode - that “red lines” on autonomous weapons and mass surveillance are non-negotiable regardless of competitive pressure - is consistent with what actually happened. That consistency is rarer than it should be.
What The Numbers Say
The commercial results are striking. Revenue grew roughly 80-fold in Q1 2026, well ahead of the tenfold growth Anthropic had internally projected. By May 2026, the annualised revenue run rate had reached approximately $47 billion. The $965 billion valuation - reached after a $65 billion Series H in May 2026 - pushed Anthropic past OpenAI’s $852 billion for the first time.
The growth rate is harder to contextualise than the absolute number. An 80x jump in a single quarter is not normal revenue growth. It reflects enterprise adoption moving from pilot to production at a speed that infrastructure rarely supports. Claude’s enterprise positioning - the predictability and controllability argument - appears to be resonating with exactly the companies that can afford to spend serious money on AI tooling.
Dario revealed in the episode that he has only one direct report: his chief of staff. Daniela runs all operations as President. The division is unusually clean for a co-founder pair. It is also the kind of structure that scales - one person focused on external strategy and research direction, one person running everything internal. The 80x revenue quarter suggests the operations side is working.
The IPO Question
Anthropic confidentially filed its S-1 with regulators on June 2, 2026. Confidential means no public financial disclosure yet and no committed timeline. It could be summer 2026, autumn, or withdrawn entirely.
The signal matters regardless of timing. An S-1 filing means the company has reached the point where public markets become a plausible next step. At $965 billion, Anthropic would be one of the largest technology IPOs in history if it proceeds. The more interesting question is what happens to the safety-first mission once quarterly earnings calls become part of the governance structure. Public companies face different pressures than private ones. The public benefit corporation structure helps, but it has never been tested at this scale.
Dario’s 25% estimate of catastrophic risk from advanced AI - given publicly several times and not retreated from even as the commercial pressure to soften it has grown - will be an interesting thing to say on a public market roadshow. I suspect he will say it anyway.
My Take
The Bloomberg documentary is worth watching on its own terms. But the reason it matters beyond the access piece is what it documents: a lab that started with an unfashionable argument (safety is the point, not a constraint) is now the most commercially valuable AI company in the world.
That outcome was not inevitable. The Pentagon dispute could have been resolved by removing the guardrails. The safety language could have softened as the revenue grew. Neither happened. The safety-as-strategy thesis - that enterprise customers value predictability more than raw capability as AI moves into production - is being validated in real money.
Whether Anthropic can hold that position at IPO scale, under public market pressure, with the Claude model family releasing faster than ever - that is the genuinely open question. The Bloomberg episode is a document from the moment just before we find out.
Dario said he is “deeply uncomfortable with these decisions being made by a few companies, by a few people.” He runs one of those companies. The discomfort is honest. It does not resolve the tension. But the Pentagon story suggests that discomfort is doing some real work on the inside.