He Never Ran a Company. Now He Runs Anthropic | Fortune Daily

0:00Anthropic co-founder and CEO Dario Amodei is a physicist turned neuroscientist who had never run a company before, but now he's operating one of the most consequential businesses on the planet, and that's Anthropic. The artificial intelligence titan has tension with the White House, with tech leaders, and perhaps also a little bit with himself. In 2020, he helped set off the entire data center boom that we are currently living through. And now he's the loudest voice warning about what happens next. And he's taking his company public very soon with a prospectus that lists existential risks to humanity.

0:43Fortune AI editor Jeremy Kahn has spent years in the room with Amadei. He joins us to share who he actually is. From New York, I'm Ellie Austin, and this is Fortune Daily.

1:00Jeremy, it's great to see you from London. Thank you for joining us. So you are definitely the man to talk to about Dario Amadei because you have met him, you have interviewed him. And for a very long time, people who followed casual AI developments knew that there were some big characters on the scene. There was Elon Musk. There was Sam Altman. But Dario Amadei wasn't necessarily as prolific a character. A few existential threats and a very important SNL sketch later, that has changed. And he is now never out of the headlines. So tell us who Dario is and why he matters now more than ever. So Dario Amadei is important now because he's the CEO of Anthropic, which is one of the leading companies producing AI models.

1:46It's also a company that's about to have a very big initial public offering. It is about to go public at a valuation. Some people have estimated it may go public at a valuation as high as$2 trillion. And you have spent time with him. So you wrote a cover story on him that also featured his sister, Daniela, who is the president of Anthropic. and I believe you also interviewed him at a Fortune event live on stage a couple of years ago as well. Tell me what it's like to be in a room with him. Paint us a picture of the person. Yeah, so I've interviewed Dario quite a number of times, both at Anthropics offices and, as you say, on stage at some events.

2:22He's quite an interesting character. He's a scientist by training, and he kind of appears as a scientist. And sometimes, you know, people sort of say he looks a bit like a mad scientist out of central casting. He's got this very curly hair. He kind of twirls ringlets of his hair when he talks, often if he's thinking of an answer. He's a thoughtful guy. He definitely sort of tries, you know, thinks about your questions. He isn't one of these CEOs who just sort of repeats talking points all the time. Yeah, he comes off as very personable. He's one of these people who has, he speaks quickly. And I think he thinks even faster because he actually sort of stumbles over his words a little bit to always get them out.

3:05So he stutters a little bit as he talks. But yeah, he comes off as a very kind of personable guy. And it kind of appears as a scientist, I'd say. And he has had a really interesting background. So he was a physicist, he was a neuroscientist, as you said, then an AI researcher. And he told do I remember in our Fortune cover story, I believe it was December of last year, December 2025, that he had never run a company before Anthropic and he knew nothing about business. So how does someone who claims those two facts as their credentials end up running a company like Anthropic that is so prolific in the culture and worth so much?

3:45Yeah, it's a good question. I mean, he ended up running this company because he really led this group of AI researchers from OpenAI and employees from OpenAI that broke away to found Anthropic back in 2021. And he, yeah, he was, he obviously has, there is some charisma about him, I have to say. I mean, he's not a character completely without charisma. And he kind of led this group that had this exodus from OpenAI. Why did they leave? What was their reason for the exodus? Yeah, so that's a good question. They left essentially because they were concerned that OpenAI wasn't paying enough attention to kind of AI safety.

4:24They were worried that there were risks and that OpenAI was kind of pursuing a profit and kind of a commercial approach and putting those risks to some extent on the back burner. But there were also, frankly, personality clashes. He did not get along very well with Sam Altman, OpenAI CEO. He did not get along very well with Greg Brockman, OpenAI's president. There were clashes between Brockman and Dario. There were clashes between Brockman and Daniela Amadei, his sister. They felt like in some of these cases that Altman was kind of duplicitous in the way he'd been dealing with various employees, including Amadei.

5:05So yeah, basically they got very fed up with the situation at OpenAI. And as a result, they decided that this group was going to leave. They didn't really feel like Altman was the right guy to be in charge of this super powerful technology. And Dario is often mentioned as one of the people who contributed significantly to the data center boom that we are seeing right now. Tell me what role he played in that. Yeah. So Dario, when he was back at OpenAI, he wrote two very kind of important papers. One was called The Great Blob of Compute that came out in 2017. and that talked about the importance of compute to AI progress.

5:44And then he wrote another paper at OpenAI in 2020 that laid out what are sometimes referred to as the AI, you know, scaling laws. And these are not like physical laws, but this was an observation, an empirical observation that basically there was a, if you increase the amount of computing power you use to train an AI model, it would increase by about an equivalent amount of performance. So if you gave it 10 times more compute, it would be 10 times more powerful. And there was a sort of set ratio between the amount of compute you were giving and the amount of power or capability you were getting out of the model.

6:19And it almost didn't matter what else you did. The more compute you gave the model, that was the biggest determinant of its performance. And that led to the building of bigger and bigger models and training them on more and more data, on more and more compute. So that kind of set off the race that's still going on for data setter capacity. Really interesting. We mentioned that Dario, he runs the company with his sister, Daniela, but there are also five other co-founders. I'm really interested in the dynamic between him and Daniela. You've profiled them both. So take us inside how that works between the two siblings.

6:53Yeah, so I think they're very close and they both say that they spend quite a lot of time together, both at work and sort of outside of work. But they've split their roles so that Dario is the classic CEO. He handles a lot of the outside role, sort of interfacing between Anthropic and the outside world. Whereas Daniela is much more like, her title's president and she functions more like a chief operating officer. She handles a lot of the internal dynamics within Anthropic. So she's dealing with personnel issues. She also has been in charge of a lot of their sort of policy efforts. So how they present themselves to government, what they lobby for.

7:34She's had some direct say in a lot of that. And I want to talk about Anthropik's relationship with government because it has really been in the headlines in recent months. So the White House AI czar has called Dario a doomer, and he's accused him of regulatory capture. Is there any truth to that? Well, it's interesting. I think most people would probably characterize Dario as a quote-unquote a doomer. It depends what you mean by that term. Dario himself says he's not a doomer, that he's actually more optimistic than not about how AI is going to go. But he does see that there's quite a significant potential for AI to go badly, and if it goes badly, for those harms to be very large.

8:15So to that extent, I guess he would be characterized as a doomer. I think he would say that there are some people who are even more concerned about existential risk from AI than he is, and who thinks that the only responsible thing to do should be to stop working on the technology altogether. He's not at that extreme, of course. Interested in what motivates him because he's not having a particularly easy time of it. He has not chosen an easy job and position for himself. And he does have this unconventional route to running Anthropic that we mentioned earlier. I was reading a lot about him in preparation for speaking to you, Jeremy, and I understand that his father's death quite early in his life plays quite a pivotal role in him wanting to develop technology.

9:01What do you know from your reporting about that? Yeah, I mean, he often talks about this. He says that, and it's one of the reasons he says he's not a doomer when it comes to AI, because he really believes in the potential of AI to kind of advance science, and in particular to advance medicine. He always says that his father died of a disease that, you know, a few years later would have been cured if that treatment could have come to market much sooner. And he really sees AI as a way to hopefully bring a lot of treatments forward, cure disease, so more people hopefully don't have to suffer the deaths of loved ones.

9:36And I think he does kind of believe that very passionately. So that's definitely had an effect on him. I'm also sure that the closeness he has with his sister is, to some extent, the result of having lost their father at a fairly young age. And he has a bit of a mystique around him personally, but the Wall Street Journal recently published a piece about his wife, a woman called Cammie Clark, who formerly dated Eric Schmidt. And she's also very integrated in the tech world, but she's quite a difficult figure to get details about. What do you know about her? Yeah, I mean, so it's very interesting.

10:09Cammie Clark is a figure who's kind of been around Silicon Valley for a long time. She tried to run a number of different startups, including one of the more recent ones was going to be a company that was a sort of a feminist pornography filmmaking company. And she tried to solicit funds famously, which come out in the Epstein files from Jeffrey Epstein. She tried to get funding from him. It does not appear that he ever provided any funding to her. And actually, that business did not last. It went bankrupt. And then she kind of, yes, as you said, she dated Eric Schmidt. She was involved in trying to sort of direct some of Eric Schmidt's AI investments for a while through various proposals.

10:50And then she met Dario, and they've actually been dating ever since 2014. She's now very involved as a kind of, you know, she doesn't have any official role at Anthropic, but a lot of people at Anthropic say she does have quite a bit of influence in the company through Dario. She accompanies Dario on almost all of his trips. She, I remember when Dario spoke at Fortune Brainstorm Tech a few years ago, Cammie was there with him behind, in the green room. And she was kind of acting, I think she kind of calms him down. As we've said, he's not someone who is sort of a natural CEO figure. So I think he relies on her for, you know, to sort of get him ready for some of these public appearances, make sure he looks okay, make sure he's kind of in the right mental frame to speak.

11:40But she also apparently has had quite a role in advising the company on some other decisions, including some of these policy decisions with the Trump administration. People sometimes refer to her as a sort of hidden eighth co-founder of the company. Yeah, so she's interesting. How do people feel about that? Because people at the company feel about that? Because, you know, There's potentially a significant conflict of interest if one of the co-founder's wife is very involved in the day-to-day runnings of the company. Yeah, so it's interesting. When I've talked to people that haven't said much about this, the Wall Street Journal reported that there was some resentment within the company about her role because she does not have an official position.

12:20But I do think she's fairly aligned with the other major executives there in terms of what she advocates for, what she kind of believes in. So, yeah, I don't know. It'll be interesting to see how this plays out once Anthropic is a public company. On that note, what do you think the next chapter in the Dario story is? You followed him so closely for years. Where does he go next? Well, I mean, I think, you know, look, I don't think he's going anywhere in terms of I think he's he's in the CEO role and the way they've actually created the corporate governance structure of Anthropic. The co-founders collectively have 50.1 percent of the voting power.

13:00So they have quite a bit of say over what happens at the board level. Now, they don't have the complete say because Anthropic also has this long-term benefit trust, which actually gets to appoint the majority of directors. But in terms of the other directors on the board, they pretty much have to do what the co-founders say. It is possible they could conceive that maybe in the future there'd be some clash between the long-term benefit trust and the co-founders. And they could potentially, they do have the ability to potentially fire Dario at some point. But I think it's highly unlikely that that would happen.

13:37Yeah, no, I think he really wants to be here to see this technology develop. He's one of these AI founders who really believes that we will achieve something called AGI or artificial general intelligence. This is AI that's as smart as any individual human. In fact, he thinks that we can go beyond that to something that's often called artificial superintelligence, where you have AI that would be more intelligent than all of humanity's kind of collective intelligence. And I think he really thinks that Anthropic can be the company that kind of midwifes this technology into the world. And he wants to be the person who's leading the company to that achievement.

14:12Jeremy, he's a fascinating guy. Thank you for being our Dario expert. We really appreciate it. And we will talk to you soon. Absolutely. Thanks for having me, Ali. Now, some more Amadei, but from a different angle. Fortune's Ali Garfinkel takes us inside how investors think of the Anthropic co-founder. And we talk about whether his intellectual doom, in inverted commas, is a liability or actually the pitch itself. Ali, thank you so much for being here. We wanted to talk to you about Dario. We've just got a really interesting picture from our colleague, Jeremy Kahn, who spent time with him and profiled him about Dario the man and what motivates him.

14:51But I want to come to you as our in-house VC expert for a sense of how investors view Dario at this specific moment. So he's intellectual, but he's also cautious at the same time as working in the fastest moving industry that we've probably ever seen. So tell me what's the verdict right now from investors? So I think it's important to consider the trajectory.

15:15Allie Garfinkle: You know, Dario started off, I think, sort of generally and uniformly, widely being perceived as kind of this philosopher upstart. Think about all of the sort of discourse around Anthropic as being the most moral AI, the most responsible AI. Dario totally personified that, and I think at the very beginning that was the general consensus, if you will. Now, I think in terms of the verdict right now, I mean, the jury is still out on all of this, it's important to say. He's gotten a lot more divisive, Ellie. I think part of it is because of the kerfuffle around the Pentagon, but also just the trajectory that Anthropic has been on.

15:50Allie Garfinkle: When you're growing that fast and that much is changing, inevitably you're going to rub some people the wrong way. And let's talk about the Pentagon for a second. So Dario has a contentious relationship right now. Contentious is an understatement. To put it lightly with President Trump, how big of a flag, and we alluded to why in the conversation with Jeremy, me. But how big of a flag is that tension for the private market? It's a pretty big flag. And again, it's divisive. So the thing to consider is if you're somebody who is an investor in Anthropic, what investors in Anthropic will tell me is they'll say, Dario is so principled.

16:26Allie Garfinkle: He's so direct. He's so honest. But if you're somebody who is critical of how he's managed his relationship with the Pentagon and President Trump, you'll say, why is he doing this? Why is he creating so much friction and so much tension? Why isn't he sending someone else to negotiate? That actually at various points has happened. And frankly, it seems like the results have been better. And it is the sort of thing that, especially if you talk to defense folks, they'll sort of say, who do you think you are that you have the right to tell the Pentagon how you do or don't use your technology, which I think is a reasonably widespread view on the defense side.

16:57Allie Garfinkle: Interesting. And so Anthropic in its S1 lists Dario himself and his reputation as a risk factor, which is fascinating to read. What does that mean in practice? And have you seen a company do that before with its co-founder and CEO? I don't think so. I was actually just thinking about it as you were saying it. And I think it has definitely been implied, certainly, but I don't know if I've ever seen it stated quite so directly. And it does sort of get into sort of questions about how much of the AI game is perception, how much this is about cults of personality. Like on the day that the OpenAIS one should drop, will we see a similar disclosure around Sam Altman as a risk factor and his reputation, maybe.

17:39Allie Garfinkle: I think it's one of those things where, safe to say, pretty anomalous. Like, I would have to go look back at SpaceX. I'm pretty sure Elon Musk's reputation didn't come into the equation there. I also think that one of the things that's sort of important to say is this is a perception game as much as anything else. And what do you mean by that? So if you look at those financials from that S1, what you want to see will determine what you see. You will either see a company that is at the very beginning of a generational run, and that$80 billion loss, I'm sorry,$81 billion loss is a sort of drop in the bucket compared to what will happen.

18:14Allie Garfinkle: Or you see a business that is going into the public markets with the greatest net loss maybe anyone's ever reported on an S1, as far as I can tell, and sit back there and say no. So it's one of those things that I think that at a certain point, Dario's reputation will exist completely in tandem with Anthropic, but also I would say at a certain point, Dario is Anthropic, actually. And over the past few months, he's very much been at the forefront of these AI doomsday conversations and been a proponent of regulation. How does that sit with investors? So a lot of investors do like it. Investors in Anthropic, it should be said, who I speak to do like it.

18:58Allie Garfinkle: They sort of say, yes, the long game here is being a hat on regulation, being seen as a trusted voice. And they do still see Dario as that. Even if it slows the company down? Even if it slows the company down is what they will tell me. I do think there has been some sort of traction on this idea of slowing down. Whether it has actually happened is another question. Whether it's actually happened at OpenAI or Anthropic is an open question I have. Definitely. But all of this to say, I do think investors, at the very least, they'll toe the line. Ali, thank you. Stay there because there's somebody else I want to ask you about in a second.

19:35Allie Garfinkle: Where does venture unequivocally belong in the process of making cancer non-lethal? Strangely, it's really important, actually. Which, you know, I never would have thought. But the funding to get things from idea into people is so crucial because the failure rates are so high. The risk of science is so great that very few people actually have the financial wherewithal and the scientific acumen to really do this well. And luckily in America right now, there's two dozen great biotech venture capital groups who I'm fortunate enough to work with a lot and who've taught me so much about this field.

20:12And we have a great robust ecosystem here. And that is something that really no other country in the world has. and it actually makes this not only a tool that we can really replicate, but it has created an ecosystem that allows me to just plug in and fund things really early and have partners down the road who are willing to then meet me when things are successful. You sat down with Reed Jobs for your term sheet show, your term sheet vodcast. For people who don't know him, give us a bit of a biography. Who is he?

20:45Allie Garfinkle: So Reed Jobs is one of Steve Jobs' children. He's his son. And Reed, over the last few years, has been building a VC firm that does sort of two things. Traditionally invents in sort of cancer care, cancer outcomes, and offers grants. And it's one of those things where it's kind of a unique structure. And he actually has sort of taken his father's death and sort of made it his life's mission. His father died of pancreatic cancer. His father died of pancreatic cancer, Steve Jobs. And he's made it his mission to make cancer non-lethal. The phrasing on that is very specific. Yeah, non-lethal because Jobs says that he doesn't think it's realistic to cure cancer.

21:21So what is he aiming for when it comes to treatment?

21:24Allie Garfinkle: So there are, I think, a couple of different outcomes. So he's sort of investing kind of across the stack, if you will, as they would say in tech. He's invested in all sorts of companies, including really early science companies to AI drug discovery companies. And the non-lethal distinction, I think, is really important. Anyone I talk to sort of says cancer is intractable. It's not like you're going to cure the flu, for example. But eventually we can get to the place where our ability to detect cancer improves wildly. Pancreatic cancer, for example, famously difficult to detect. So he's definitely invested in detection capabilities.

21:59Allie Garfinkle: And also improving outcomes. Chemotherapy actually was wildly different even 30 years ago. And it's one of those things that if you're somebody like Reid, what you sort of say is actually the treatments can be improved with better science. And is he focused on specific types of cancer, or is he thinking across the board? He's thinking across the board. He definitely has a couple of specific interests. Actually, this is maybe kind of weird, but I actually have a genetic mutation that makes me slightly predisposed to certain types of cancers. It's called Lynch syndrome. He actually used to study that in school, so we ended up talking about it.

22:33Allie Garfinkle: He's, of course, also interested in pancreatic cancer, as you might imagine. Did he tell you anything about your predisposition that you didn't know? No, he actually...we had a whole conversation about it because there are many sort of variations of it. I won't get into the various genetic mutations, but the specific one I have, he actually had been studying. And it was one of those things where he had a conversation. He was like, actually, maybe you should go look at this cancer vaccine and follow some of the trajectories about all of how this is developing. It was honestly really interesting.

22:58Allie Garfinkle: He is...I've interviewed a lot of VCs who are sort of purporting to be in the biotech or cancer space. He was definitely the most scientifically rigorous to the point where there are times where I was like, I'm a tech journalist. I don't know where we're going here, but let's do it. And it's really interesting because to use the modern vernacular, he is in many ways a Nepo baby. And he could have sat back for the rest of his life and never had to earn another penny. But he's chosen this very complex, very significant pursuit for himself. Give us a sense of what that says about him and who he was when you spent time with him.

23:33People are fascinated by Steve Jobs, understandably. How does it feel to sit down opposite his son?

23:40Allie Garfinkle: So incredibly normal. I will say he's somebody who I vastly enjoyed talking to because it was so intellectually rigorous. And it's the kind of thing where people are very caught up. Like I was looking at some of the comments. Everyone's like, oh my God, he looks like his dad. I'm like, God, I'm sure this guy gets this all the time. But I will say one of the things I did find really interesting is to your point, this is somebody who could have chosen a lot of lives, right? And I always am interested in people's through lines, past as prologue. and his father passing away from cancer that was undetectable for a long time clearly affected how he saw the world.

24:16Allie Garfinkle: And I think growing up in sort of the milieu of tech, I wonder if he took away the right lessons where he sort of sat there and said, what if it doesn't have to be this way? What does it mean to invest towards that? It's funny, he is a VC, but I don't think of him necessarily as a traditional VC. And maybe that's something that sort of his Nepo baby status has given him the ability to navigate in a different way. But it was very interesting to sit across from someone who has truly dedicated their lives to one purpose. And talking in three lines, we were just speaking earlier about how Dario Amadei also motivated by the premature death of his father.

24:52Ali, where can we watch your full interview with Reid? You can watch it on YouTube, Term Street Podcast. Thank you so much, Ali. A delight to see you. Delight to see you, Ali. That's it for Fortune Daily. I'm Ali Austin. please remember to like the show and subscribe to our channel because it helps other people find what we're doing. In the meantime, I'll see you tomorrow.