Anthropic: The $965 Billion AI Juggernaut Navigates Pentagon Standoff and Questions About Job Destruction
Bloomberg interview with Dario and Daniela Amodei, June 10, 2026
Anthropic CEO Dario Amodei has placed the odds of civilization collapse from AI somewhere between 10% and 25%, yet his company continues to race ahead, now valued at nearly a trillion dollars and generating profitable revenue for the first time. In a rare extended interview with Bloomberg's Emily Chang, the OpenAI defector who co-founded Anthropic with his sister Daniela defended the company's contradictory stance of simultaneously warning about existential risks while shipping products at breakneck speed.
The Pentagon Ban and Defense Dilemma
The most significant recent development emerged in early 2026 when the Department of Defense demanded Anthropic allow full use of Claude without guardrails. The company refused, drawing red lines against mass surveillance and autonomous weapons systems, which led to Anthropic being blacklisted from the Pentagon despite holding a $200 million contract won in 2025 alongside OpenAI, xAI and Google. Defense Secretary Pete Hegseth called Dario an "ideological lunatic," but Amodei appears unfazed by the characterization.
"This is more a debate about what the proper use of AI by the government is," Dario told Chang. "AI is an emerging new technology. We don't understand the ways in which it's reliable or unreliable." He framed the standoff as establishing precedent for appropriate use cases, acknowledging that "we were willing to risk the future of our company to limit how these models are used."
The company's position has created awkward contradictions. While banned from the Pentagon, Anthropic has been working with defense contractor Palantir since 2024, and Bloomberg has reported that Claude is being used by the US military for AI-assisted targeting through Palantir's Maven Smart system. When Chang pressed Amodei on whether Claude played a role in a February strike that reportedly hit an Iranian girls' school, killing more than 150 people, most of them children, Amodei acknowledged the tragedy but said "this is a use case that doesn't even violate our red lines."
His logic centers on maintaining human decision-making authority. "The principle that was obeyed here is a human makes the final decision," he explained. "If this isn't an illustration why that principle is so important, I don't know what is." He defended partnering with the defense establishment as necessary in a world where "Russia invading Ukraine" and "the risk of China invading Taiwan" require the US to maintain technological superiority over what he calls "a resurgent authoritarian block."
Mythos and the Cybersecurity Superweapon
Perhaps the most alarming revelation in the interview concerned a new AI model called Mythos that Anthropic developed but chose not to fully release. The model identified thousands of cybersecurity vulnerabilities across major operating systems, prompting some early recipients to call it "a super weapon" and plead with Anthropic not to release it publicly. "This is a super weapon, you should have to own a gun license to use it. Please don't release this," Amodei recalled companies telling him.
Through an initiative called Project Glasswing, Anthropic gave select organizations access to Mythos, including federal agencies like the National Security Administration, despite the company's Pentagon blacklisting. Dario described it as "a particularly large jump" in the models' ability to find vulnerabilities, expressing surprise at the magnitude of the capability increase.
When challenged on whether Anthropic should have the power to decide who gets access to such technology, Daniela defended the decision as grounded in "a very specific concern around cybersecurity." She acknowledged the complexity but insisted "we've tried to be as publicly open as possible to say we're trying our best to make this decision well, but we might not do it perfectly." Dario pushed back against characterizations that the restricted release was marketing theater, noting "we have suffered enormously commercially from not releasing this model."
The Job Destruction Question and Revenue Explosion
Anthropic's revenue has skyrocketed, with the company reporting first-quarter annualized growth of 80x year-over-year and API volume up nearly 17x. The company achieved profitability for the first time, driven largely by enterprise-focused products like Claude Code and Claude Cowork that automate software engineering and knowledge work.
Dario has been unusually direct about job displacement, previously estimating that AI could eliminate half of all entry-level white-collar jobs within one to five years. When Chang asked if that prediction still holds a year later, he responded: "I don't know exactly, but I'm still pretty concerned. I'm still the same order of concerns." He described an uncomfortable future where "AI could have this very unusual combination of very fast GDP growth and high unemployment, or at least underemployment or low wage jobs."
Boris Cherny, the engineer behind Claude Code and Cowork, exemplified this shift. He reports that Claude now writes "almost all" the code on his team, and for the past six months, 100% of his own code has been AI-generated. "The work of engineering has just completely changed," Cherny said. "I feel like I suddenly have superpowers." Yet he acknowledged the dark side: "The concerning part is they can do things we couldn't dream of and timelines we couldn't ever expect. You really have to be prepared because they will really be a thousand times more productive."
Dario vigorously defended himself against accusations from figures like Jensen Huang that his warnings about job loss amount to "scaring people" or "doom marketing that benefits Anthropic." He called such criticisms "cheap marketing" that's "part of the disease of Silicon Valley" caught up in "this social media world of three seconds." He insisted his essays like "The Adolescence of Technology" lay out detailed economic policy responses including taxation and identifying new job categories, but that social media clips distort his nuanced position.
For potential solutions, he pointed to expansion of the physical world economy, human-centered relationship jobs, and roles directing AI systems, though he admitted uncertainty about "how thin versus how thick" that last category will be. Daniela expressed more optimism, using medicine as an example where AI might handle diagnostics while humans focus on bedside manner and physical examination.
The Enterprise Bet and SaaS Apocalypse
Anthropic's strategic focus on enterprise rather than consumer applications appears to have been both a values and business decision. Dario explained the logic: "If you pick a business model that fundamentally conflicts with your values, you're gonna have a hard time. Either you betray your own values or you become irrelevant." He contrasted social media's engagement-driven, advertising-based model with enterprise work in biotech, pharma, and energy that aligns with using "AI to cure diseases that we couldn't cure before" and "make energy cheaper and more efficient."
The release of Claude Cowork triggered what traders dubbed the "SaaSpocalypse," wiping $285 billion in market value from software stocks virtually overnight as markets grappled with how much traditional software gets replaced by AI. Dario acknowledged that "some of them may go down in value, some of them may even go out of business if they don't adapt in the right way," but argued the overall software industry will expand even as individual companies face existential threats.
The OpenAI Split and Founding Story
The Amodei siblings left OpenAI in 2021 along with five other co-founders, all of whom remain at Anthropic today, a rarity in Silicon Valley. When pressed on the reasons for leaving, Dario was more direct than in past interviews about trust issues with Sam Altman. "When you feel that you can't trust someone, when you feel that their values are not what they say they are, when you feel that they're not honest, that makes it very hard to continue to work with a company," he said. "Why argue with someone when you don't have the same vision and you don't trust them? The way to resolve it is you go off and do your thing."
At OpenAI, Dario developed the concept of scaling laws, predicting that large language models would improve simply by adding more data and computing power. "At that point in time, not a lot of people believed scale up is the way that these models are gonna get smarter and better," he recalled. That approach helped supercharge OpenAI's models and paved the way for ChatGPT, but the philosophical differences proved irreconcilable.
The founding team would meet during the pandemic at Precita Park in San Francisco, pulling up chairs on the grass to discuss what they were building. From the start, Anthropic positioned itself as the safety-conscious alternative, with Claude trained to follow a "constitution" of principles. When asked whose values go into Claude, Daniela pointed to "founding documents in human history, like the UN Declaration of Human Rights," and said the company has begun conversations with religious leaders about "core values that are consistent across religions."
Regulatory Stance and Government Involvement
Dario took pointed shots at both ends of the political spectrum on AI regulation. He criticized a "particular group of people in the tech world in Silicon Valley" who started with extreme anti-regulatory positions arguing that even "transparency around this technology, even export controls" would "apocalyptically destroy our potential to create the technology," but then "as soon as they see the first real danger" started talking about "nationalization and the government should just seize it."
"You're yo-yoing from the most extreme anti-regulatory to this completely communist, the government should grab it all," he said. "We need a more sensible, moderate approach." His preferred framework includes "basic regulation of the technology" with "required pre-release testing and auditing of the models," an approach the initial Trump administration rejected when President Trump dismantled Biden's AI executive order on his first day back in office.
The White House position, articulated by former AI and crypto czar David Sachs, favored a hands-off approach with Sachs warning that "excessive regulation of the AI sector could kill a transformative industry just as it's taking off." However, the Mythos development and its national security implications appear to have shifted the administration's thinking toward wanting to gatekeep the most powerful AI.
Dario also criticized chipmakers for selling AI chips to China, calling it "a bit like selling nuclear weapons to North Korea." He said he's been "very outspoken about the need for export controls on chips to China" because "it would be really bad for America, for the state of democracy in the world for China to be ahead in AI capabilities." He acknowledged his chip partners "obviously don't agree with that view, but it hasn't stopped me from saying it."
The Sibling Dynamic and Company Culture
The interview revealed an unusual corporate structure where Daniela runs day-to-day operations with the entire leadership team reporting to her, while Dario has no direct reports. "It's incredibly freeing," Dario said. "It lets me do all the things that I do much more easily than I would otherwise." When Chang joked "she does all the work," Dario deflected: "If you had to go through the things I had to go through during Department of War..."
Daniela described their complementary backgrounds, with Dario taking calculus in middle school and math classes at UC Berkeley during high school while focusing on science, while she was "more into reading and arts." They lived together in San Francisco along with Daniela's husband Holden Karnofsky before joining OpenAI in 2016 and 2018 respectively. "Dario and I, we've always been really close since we were little, but I think we always really wanted to do something big together," Daniela said.
Dario emphasized internal communication as crucial to handling pressure, holding uncensored two-week all-hands meetings where he talks "for an hour about what's on my mind, what's going on in the industry, what's going on in the outside world." This approach creates what he calls "3,000 people who are on the same page as me," an "incredible amplifier" that means "I never feel like I'm alone" when confronting external challenges.
Existential Risk and the Oppenheimer Question
When asked about parallels to Oppenheimer given his love of "The Making of the Atomic Bomb," Dario identified instead with Leo Szilard "who first had the idea that there could be a chain reaction." He views Oppenheimer as "a failure case, as what should not happen," arguing that "we're not gonna get through this with larger than life personalities or figures who try and be at the center of everything."
On the 10% to 25% probability of civilization collapse he's previously cited, Dario said that risk "comes from the very straightforward recipe of the technology, the existence of many countries in the world, the existence of many companies within an economy." He drew an analogy to airlines: "It can both be the case that your airline company is 10 times safer than all the other airline companies. But if someone comes and asks you, can you guarantee that your airplane will never crash?" When Chang noted that no one would board a plane with a 25% crash risk, Dario replied: "25% is too high. We're trying to make that probability much, much lower. That is the goal."
On whether something Anthropic builds could cause that collapse, he said "I certainly hope not" but acknowledged "it's never gonna be zero." He framed Anthropic's role as trying to be the safest option in an inevitable race: "Half of what we do within the company is try and reduce the risk as much as we can." The underlying premise is that someone will build advanced AI regardless, so the question becomes whether the builders prioritize safety.
Dario's stress relief involves occasional weekends playing video games, sometimes with Daniela, and trips to Italy where he and his wife have a horse named Calypso. "She doesn't know about any of this, she's just a happy horse," he said, describing the experience of sitting with the animal as a form of zen.
When asked why people should trust Anthropic given Silicon Valley's credibility crisis, Dario acknowledged that "starting from a position of distrust is pretty rational" and that "Silicon Valley has lost a lot of the world's trust and has to re-earn it." His pitch: "The message we're trying to send is, we're actually different and that has to be earned in things that we actually do." Whether that proves true will depend on how Anthropic navigates the escalating tension between its safety mission and the commercial and political pressures of building what may be the most consequential technology of the century.