Cameron Costa (2:07)

I'm CNBC producer Cameron Costa. Squawkpod reports from Davos 2026 Anthropic CEO Dario Amode begins right now. Davos, Switzerland. Snowy, chilly and this time of year also filled with leaders from every field. Banking, government and duh, artificial intelligence. Joe Kernan, Becky Quick and Andrew Ross Sorkin spoke to many execs this week, including OpenAI's CFO Sarah Fryer. You can catch that interview on your podcast feed if you scroll down. They also spoke to another giant in the AI space, Anthropic founder Dario Amodei. Anthropic's Claude model is up there with ChatGPT and Gemini in terms of popularity, but Amode says it has an edge. Here's Andrew on our outdoor set in Davos.

Andrew Ross Sorkin (3:11)

And our next guest is on the front lines of the AI revolution. Joining us right now is Anthropic's co founder and CEO Dario Mode. Great to see you, sir.

Dario Amodei (3:19)

Thanks for having me, Andrew.

Andrew Ross Sorkin (3:20)

So here we are. You just came out with a new model last week. A lot of people are buzzing about that here. I think it would be very helpful for the audience to understand how you see what you're doing, what the folks at Google are now doing, what OpenAI is doing, what you think XAI is doing, sort of what that actually even looks like anymore. Because it does seem like everybody's trying now all of a sudden to differentiate in different ways.

Dario Amodei (3:44)

That's right. So you know, I think we've, we've picked a path that's quite different from the other players. You know, Anthropic, since the beginning, you know, Anthropic has thought in terms of safety and reliability of AI systems. And one of the things we realized is that that was very synergistic with working with enterprises as compared to consumers. Because enterprises really, they value the reliability, they value the trust, they value the long term relationship with customers. And so we decided to pursue the enterprise business. We have, you know, we have consumer products, but they're not our major focus. It's maybe 80, 20 on, it's maybe 80, 20 on the enterprise side. And so that, that really allows us to proceed in a different way. Right. We don't have to do ads, we don't have to focus on short form video, we don't have to focus on maximizing engagement. Instead we actually just focus on making our models as smart and as capable as possible. Because that's what enterprises are looking for. The, they're looking for capability, they're looking for intelligence, they're looking for optimizing complex business processes across things like finance, coding, pharmaceutical companies, manufacturing, energy. Right. All the things that drive the economy. So we think of the work we do as most directly connected to directly driving GDP and not all this kind of indirect or complex like maximizing engagement.

Andrew Ross Sorkin (5:04)

But is the complex stuff video and images and all of that and ultimately do you think that those models will be able do what your models do?

Dario Amodei (5:12)

So you know, I actually think there's a divergence where if you're optimizing, if you're optimizing for consumers, right, you're trying to be as engaging as possible, you're trying to be as entertaining as possible. I actually don't think some of these other models have enough incentive to get smarter, have enough incentive to become capable of more tasks across the board. And we've been able to focus on that more.

Becky Quick (5:35)

I mean, that probably responds to some, it's probably music to the ears of some investors. We've spoken with investors. Satya Nadella was just talking about this at Microsoft. You know, his point is that if we don't see a broadening of more than just the big tech companies using this AI, if we don't see a broadening of just the wealthiest nations using it, then it is indeed a bubble, that there's too much infrastructure that's being built up. But maybe that speaks to all who's going to pay for it and who's not.

Dario Amodei (6:05)

Well, you know, I mean, I think across the board we're both trying to grow our revenue as fast as possible. Our revenue grew from zero in 2023, zero to 100 million, 2024100 million to 1 billion and 20251 billion to almost 10 billion. So if we can keep that curve going or even anywhere close to it, then, then you know, the return, the investments will be justified. But I think you bring up another point that's also relevant, which is, you know, we need to make sure that kind of the technology diffuses and everyone shares in the benefits. The startups at Silicon Valley, you know, adopt, adopt this technology almost immediately. I think it's a little harder to get into the large enterprises, although we're spending an increasing amount of time on go to market work to try and drive that enterprise transformation and get enterprises to adopt it as quickly as possible. But then, you know, there's all kinds of countries in the world so you know, diffusing this technology across, across not just the developed but, but the developing world, I think that's going to be very important. And you know, we've done, we're starting to do work with folks like the Gates Foundation. Could you just estimate 10 years revenue growth? You know, it's, it's, it's, it's very, it's very hard to say.

Cameron Costa (7:11)

Should I ask?

Dario Amodei (7:12)

Yeah, anything I say is going to sound totally crazy, but if you, if you look at what AI is capable of, if you have these models that are getting more and more capable across a wide range of cognitive tasks, you know, you look at all labor in the economy, that's something like $50 trillion a year. So I could easily imagine that the revenue of the industry or even single companies, if it's even 10% of that, could be $5 trillion a year. Now, that's something we haven't seen in the history of the world that creates all kinds of problems as well as creating all kinds of growth.

Andrew Ross Sorkin (7:45)

When you think about though, the growth and you know, there's obviously an expectation that you'll go public as potentially OpenAI and others do as well, how much more capital are you going to need to raise between now and whatever that moment is? Because there's sort of the technology piece on one end and then there's the capital raising piece. I don't know if you ever got into this thinking you were going to be a capital raiser all day long.

Dario Amodei (8:07)

Not really, no. So, you know, the way I thought about it and you know, we talked about it when we, when we spoke at, when we spoke at DealBook is there's this kind of difficult problem, this cone of uncertainty, right, which is the revenue is growing very fast. You don't know exactly how fast because the rate of, of, of enterprise diffusion is unknown. But you know, you don't know whether your revenue next year is going to be, you know, 10 billion now is going to be 20 billion, or is it going to be 100 billion next year? You don't know. You don't know anywhere between that and yet you have to buy the data center.

Becky Quick (8:38)

So what do you do?

Andrew Ross Sorkin (8:38)

60 to 80%? Well, for 10 years, starting at 2024 to 2028, 130 to 150. Which model is telling you this is Chat GPT? Don't tell him that. I thought they'd, I don't know.

Becky Quick (8:53)

I mean, what do you do when the cone of uncertainty is that wide of a spare?

Dario Amodei (8:57)

So there are a few things that are your buffer. The first thing that's your buffer is, you know, the higher your margins are, the more, you know, the, you know, the more fairway you have between the amount you have to buy and the amount of revenue you have to support. So one of the fortunate things about being in the enterprise business is that we have relatively higher margins and enterprise spending is relatively predictable in comparison to consumer spending, which is very fickle. Right. We see Google and OpenAI fighting it out over consumer spending. We have more stability in enterprise. The other thing, of course is, you know, while you're growing, that revenue is of course is of course raising money. And so, you know, we do that periodically. You know, we raise, we raise very large amounts of money and I would expect those to continue to increase, but that as time goes on, our revenue will continue to increase fast and will.

Andrew Ross Sorkin (9:44)

Justify what is the moat around the business? And the reason I ask is I always thought persistent memory we've talked about this could be the moat. In your case, it might be the data sets that you're connecting into and the cost of changing out anthropic for another model. But by the way, I did this last week. I don't think I told you this. You can go in and this is on the consumer end you can ask ChatGPT, for example, to effectively output everything it knows about you so that it can be imported into another model, in which case actually there is no moat around the business because you could port yourself from place to place. And by the way, philosophically maybe that's great since it's my information. But on the other end, I think to myself, well, what does that do to the business if your model gets better than the other one and then they go back and forth and everyone keeps flip flopping.

Dario Amodei (10:32)

So again, I think that's talking on the consumer side. On the enterprise side, I think there's a few, few moats. One is, you know, again, as I said, enterprise really, really values model intelligence, right? Being better at coding, being better at understanding finance, being better at biological discovery. So you know, we simply try to have the smartest model and there are maybe only three or four companies that are plausible for having the smartest model. So you're already talking about a relatively small group. The second thing is within that group, models specialize in different things. So anthropics models are, you know, widely, widely known to be widely considered to be the best models for coding. And so, you know, within, within the coding market, we have a moat within that area and we're trying to build moats within other areas. You know, we recently released this Claude cowork thing that tries to go beyond coding to general agentic tasks. And then I think the third thing is the thing you mentioned, which is that enterprise relationships are, you know, formed slowly over a lot of time and then are difficult to disrupt. We talked about the slow speed of economic diffusion and the need to do enterprise transformation, right, to get companies with tens of thousands of employees to actually adopt the technology. But the flip side of that is if they manage to go through that slow process with one company, it is again slow to try to do that with another company. So I think those are the basis of the bounce.

Andrew Ross Sorkin (11:54)

We got to go. But just one final question. Does all of the hand wringing and sort of splintering of our globe impact your business at all? I mean, as you're talking to people trying to sell into different countries. Is that shifting the discussion?

Dario Amodei (12:08)

No. You know, I think we're thought of as kind of an independent actor. And, you know, we just have models to sell. We're willing to sell them to businesses across the world. When policy matters come up, we do speak on them. And sometimes we agree with the administration, sometimes we disagree. But, you know, we try to be very technocratic in the way that we address these things.

Andrew Ross Sorkin (12:28)

Dari, I want to thank you for coming in. It's great to see you. Thank you. Andrew here in the cold. Thank you. Great to see you.

Commercial Narrator (12:39)

Building a portfolio with Fidelity Basket Portfolios is kind of like making a sandwich. It's as simple as picking your stocks and ETFs, sort of like your meats and other topics and managing it as one big juicy meat investment. Now that's pretty good. Learn more@fidelity.com baskets Investing involves risk, including risk of loss. Fidelity Brokerage Services, llc Member nyse, sipc.

Cameron Costa (13:09)

Thy ticket lady, Jennifer of Coolidge. Well, many thanks, good sir. Here is my Discover card. They accept Discover at Renaissance Fairs?

Dario Amodei (13:20)

Yeah, they do.

Cameron Costa (13:20)

Here. Discover is accepted at the places I love to shop. Get it with the Times. With the Times. You're playing the loot. Yeah, and it sounds pretty good, right?

Dario Amodei (13:32)

Discover is accepted at 99% of places that take credit cards nationwide, based on the February 2025 Nielsen report.

Cameron Costa (13:39)

Not every sale happens at the register. Before AT&T business Wireless, checking out customers on our mobile POS systems took too long. Basically a staring contest where everyone loses. It's crazy what people say during an awkward silence. Now transactions are done before the silence takes hold. That means I can focus on the task at hand and make an extra sail or two. Sometimes I do miss the bonding time. Sometimes.

Andrew Ross Sorkin (14:04)

AT&T business Wireless connecting changes everything.

Cameron Costa (14:14)

Thank you for listening to this special Squawk Pod reports from Davos. I encourage you to stay a while, poke around. Squawk's biggest interviews from the World Economic Forum this year live on this very feed. And they'll keep coming tomorrow and even over the weekend. Joe's full interview with President Trump, Becky's full interview with Amazon CEO Andy Jassy, Andrew's interview with Bill Gates, and so much more. Just hit that follow button and you can catch each one every time we post. Squawk Box is hosted by Joe Kernan, Becky Quick, and Andrea Ross Sorkin. Squawk Pod is produced by me, Cameron Costa and Zach Valise. Katie Kramer is in Davos, and Julie Tras, our editor, is here Stateside. Have a great day. Hey Fidelity, what's it cost to invest with the Fidelity app? Start with as little as $1 with no account fees or trade commissions on U.S. stocks and ETFs.

Dario Amodei (15:21)

Hmm.

Cameron Costa (15:22)

That's music to my ears. I can only talk.

Dario Amodei (15:27)

Investing involves risk, including risk of loss. Zero Account fees apply to retail brokerage accounts only. Sell order assessment fee not included. A limited number of ETFs are subject to a transaction based service fee of $100. See full list of Fidelity.com commissions. Fidelity Brokerage Services, LLC Member NYSE SIPC.