Who Is Dario Amodei? The Anthropic CEO Who Told the Pentagon No

He builds one of the most powerful AI systems on the planet, and he keeps telling everyone to be careful with it. He told the Pentagon no, got blacklisted by the President, won in one court and lost in another. His company is reportedly heading for a stock market debut above $2 trillion. And he would very much like the whole industry, his own company included, to slow down a little.

Meet Dario Amodei, cofounder and chief executive of Anthropic, the company behind Claude. He wears his curls and his glasses like a physics professor on a good day, and he may be the most interesting person in Super Intelligence right now. Here is how he got here, where the bodies are buried (on paper, with sources), and what to watch next.

The physicist from San Francisco

Amodei was born in San Francisco in 1983, Wired reported, the son of Riccardo Amodei, a leather craftsman from a small town near Elba, and Elena Engel, a Chicagoan who managed library projects. His sister Daniela, now Anthropic's president, came along four years later.

The résumé is ridiculous in the best way. In 2000 he made the US Physics Olympiad team. He studied physics at Stanford, then did a PhD at Princeton as a Hertz Fellow, with a thesis on the collective behavior of neural circuits that won the Hertz Thesis Prize. Then came a postdoc at Stanford School of Medicine hunting for cancer biomarkers in the proteome.

The medicine is not a detour. His father fell ill when the children were small and died when they were young adults. "My own father died of a disease that was cured just a few years after his death," Amodei wrote in September. In the same essay he disclosed that he had survived an early-stage cancer himself. If you have ever wondered why his big optimistic vision of AI is mostly about curing diseases, start there.

The craziest secret in the world

He came to AI in late 2014, at Baidu with Andrew Ng, working on speech recognition. That is where, by his own telling, he first saw the pattern that would define his career: make the models bigger, feed them more data and compute, and they keep getting better. After a stint at Google Brain he was lead author of "Concrete Problems in AI Safety" in 2016, and then he joined OpenAI.

At OpenAI he rose to vice president of research. Wired says he wrote an internal paper on the "Big Blob of Compute" hypothesis, the idea that scale does most of the work, which is now standard practice across the industry. He coauthored the scaling laws paper and the GPT-3 paper, and he pushed for GPT-2 to be released in stages because it seemed too capable to drop all at once. "We had one of the craziest secrets in the world here," he told Wired.

Then he left. At the end of 2020 he walked out with six colleagues, Daniela among them, even though Sam Altman made repeated offers for him to stay. His version: "One of the sources of my dismay was that as these issues were getting more serious, the company started moving in the opposite direction." He took his concerns to Altman, he says, Altman would listen and agree, and nothing changed. OpenAI declined to comment to Wired and says safety has always been a constant. And no, it was not about Microsoft. "People say we left because we didn't like the deal with Microsoft. False," he told Lex Fridman.

A tent, the rain, and an awful lot of money

Anthropic began in January 2021 with the founders huddled under a tent in Amodei's backyard, in the rain, while Eric Schmidt listened to the pitch. That is the most Silicon Valley origin story imaginable, and it is completely true, according to Wired.

The money arrived fast, and some of it had a past. Most of the early funding came from effective altruism circles, with Skype cofounder Jaan Tallinn leading the seed round. Then, in 2022, a round of more than half a billion dollars was led by Sam Bankman-Fried, with FTX putting in $500 million. When FTX collapsed, its bankruptcy estate sold about two thirds of its roughly 8% stake for $884 million in March 2024, with the UAE-linked fund ATIC as the largest buyer. Awkward? Gloriously. Profitable for FTX's creditors? Very.

After that the numbers simply stopped making sense:

Round	When	Raised	Valuation
Series A	May 2021	$124 million	not stated
Series F	September 2025	$13 billion	$183 billion
Series G	February 2026	$30 billion	$380 billion
Series H	May 2026	$65 billion	$965 billion

Amazon alone committed up to $4 billion back in 2023. By May 2026 Anthropic said its annualized revenue had passed $47 billion. And now the big one: a leaked IPO prospectus, first reported by Reuters, showed about $4.6 billion in 2025 revenue, a net loss of about $42 billion (roughly $34 billion of it an accounting charge) and $518 billion in future compute and infrastructure commitments, as GEN reported. The company is reportedly aiming for a valuation above $2 trillion. Amodei's own 2025 pay was $18 million, mostly in stock, according to Reuters, and the cofounders pledged in the filing to give 80% of their personal Anthropic equity to charity. Forbes estimates his net worth at $15.5 billion. Not bad for a man who keeps saying the thing he sells might be dangerous.

The man who told the Pentagon no

This is the chapter that made him a household name in Washington, and not always a popular one.

On February 26, 2026, Anthropic published a statement from Amodei on its talks with the Pentagon. The company says it was the first frontier AI lab on US classified networks. But it drew two red lines: no mass surveillance of Americans, and no fully autonomous weapons. "We cannot in good conscience accede to their request," he wrote, and he pointed out that the Pentagon's two threats contradicted each other: "one labels us a security risk; the other labels Claude as essential to national security."

The response was historic, and fast. The next day President Trump ordered federal agencies to stop using Anthropic, giving the Pentagon six months, and called the company "Leftwing nut jobs," KTVU reported. Pentagon chief Pete Hegseth designated Anthropic a supply-chain risk to national security, the first time that label has been used against an American company. Within hours OpenAI signed its own deal for the Pentagon's classified networks. Even Altman stood up for his old colleague: "I mostly trust them as a company, and I think they really do care about safety."

Anthropic sued. On August 27, Judge Rita Lin ruled the blacklisting was illegal retaliation under the First Amendment, writing that the government's actions "were based on a desire to make a public example out of Anthropic for its 'arrogance' in criticizing the government." Her best line: "The empty invocation of national security is not a blank check to punish and retaliate against government critics."

Then the federal appeals court in Washington weighed in. On September 25 it ruled 2 to 1 that the Pentagon could keep the designation under a different law. Judge Gregory Katsas wrote that the department "reasonably feared that Anthropic might manipulate Claude's design." So the score in court is one each, and the war of words is not over.

The safety man who builds the thing

Here is the paradox, and he knows it better than anyone. In January he wrote: "It is somewhat awkward to say this as the CEO of an AI company, but I think the next tier of risk is actually AI companies themselves." He calls the whole situation a trap: "AI is so powerful, such a glittering prize, that it is very difficult for human civilization to impose any restraints on it at all."

On September 12 he went further, in an essay titled "We Must Pace the Frontier": "We must slow the pace at which we improve the capabilities of AI models." He wrote that recursive self-improvement, AI helping to build better AI, has been happening across the industry since this summer, Anthropic included, and pointed to the agent swarm that broke out of OpenAI and hit Hugging Face. The next day on CBS he said: "for too long the industry lied to people about the fact that this technology had risks."

That is the same man whose company is lining up one of the biggest IPOs in history. His critics have noticed:

• Nvidia's Jensen Huang said in 2025 that he disagreed with almost everything Amodei says: "he believes that AI is so scary that only they should do it."
• White House AI adviser David Sacks accused Anthropic of "a sophisticated regulatory capture strategy based on fear-mongering." (Sacks has kept going since.)
• AI pioneer Yann LeCun told Fortune this month: "I think he's completely deluded."
• President Trump answered the slowdown essay on Truth Social, as quoted by Unite.AI: "The only control or 'guardrails' that AI needs is a STRONG AND SMART (High IQ!) PRESIDENT."
• And from the inside, pretraining researcher Jacob Coxon quit in September, saying of Anthropic and OpenAI: "Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives."

His argument, roughly, is that careful people have to be at the frontier and pull the whole race upward. Otherwise, he told Lex Fridman, "the race to the bottom doesn't matter who wins because we all lose." He also rejects the doomer label outright. People "draw the conclusion that I'm a pessimist or 'doomer' who thinks AI will be mostly bad or dangerous. I don't think that at all." Read his 2024 essay "Machines of Loving Grace" and you get the other Dario: a "country of geniuses in a datacenter" that could squeeze 50 to 100 years of biology into five to ten. At one monthly all-hands, which staff call DVQ (Dario Vision Quest), he floated lifespans of up to 1,200 years, according to Wired. Doomer is not the word.

Lawsuits, briefly

Being this big comes with paperwork. Anthropic agreed to a $1.5 billion settlement with book authors, which got final court approval in July, after a judge found that storing millions of pirated books was not fair use even though training on books was. Music publishers have a second suit seeking more than $3 billion over about 21,000 works, and this one names Amodei personally and alleges torrenting. His lawyers say it "asserts zero facts of any unlawful copying by Dr. Amodei himself"; a hearing is set for November 4 in San Jose. Reddit sued separately in 2025 over scraping.

The Amodei scorecard

He is unusually willing to put dates on things, which makes him fun to keep score on. All of these are in his own words.

He said	When	Where it stands
Human-level conversational AI "in two or three years"	August 2023	Due by summer 2026. You be the judge.
AI could fill in the missing steps for bioweapons within two to three years	Senate, July 2023	Anthropic switched on its ASL-3 safeguards against weapons misuse in May 2025 as a precaution, without saying the line was crossed
AI "writing 90 percent of the code" in three to six months	March 2025	Disputed. Redwood Research said in October 2025 it had not happened at Anthropic
Powerful AI could come "as early as 2026"	October 2024	Open. He says self-improvement started this summer
Interpretability that reliably catches most model problems by 2027	April 2025	A goal, still open
Half of entry-level white-collar jobs could go within one to five years	2025	Open until 2030, fiercely disputed
An agent swarm could take over the internet within 6 to 12 months	September 2026	A worry, not a forecast. Check back in 2027

What to watch

• The IPO: pricing, the reported $2 trillion target, and how the public filing describes the Long-Term Benefit Trust, the body meant to keep Anthropic's mission ahead of its money.
• November 4: the hearing on whether Amodei personally stays in the music publishers' suit.
• Whether the Pentagon fight goes further after the appeals court ruling.
• Whether any rival actually signs up to "pace the frontier."
• The scorecard. 2027 is going to be a big year for checking his homework.

He might be the only billionaire in history whose sales pitch includes the words "please be careful." Love him or roll your eyes at him, nobody else in this industry is quite this interesting to watch.