TMTB: Dario Amodei (Anthropic CEO) at MS TMT Key Quotes

There’s been talk about the $19B ARR number being thrown around, but that was MS in their opening remarks, not Dario stating it. He didn’t acknowledge, confirm, or deny it. Here’s the quote from the MS Host

Dario, welcome back. It’s great to see you again at this conference. Two years ago, I think I met your sister for the first time, and you just launched Claude three, I think the family, and you’re at $100 million of run rate revenue, and I had to scratch out a number a few hours ago, but I think you’re now run rating over $19 billion it’s really remarkable.

So it’s not completely clear where that number is coming from or if it’s right or not

Overall, he sounded very bullish as he should:

Key Quotes from Dario:

AI Scaling: No Wall, Accelerating Exponential

“We do not see hitting the wall. I think this year is going to have a radical acceleration that surprises everyone. Exponentials catch people off guard — there’s the old parable of the second half of the chessboard, where you have one grain of rice in the first square, two on the second, four on the fourth. By the time you get to the 64th square, you have billions or trillions of grains of rice. We’re standing on square 40 out of 64, and from square 40 to square 64, it’s going to go faster than you think — even having seen how fast it’s gone so far. I don’t think people are ready for it. I think we are on the precipice of something incredible.”

---

Code as the Leading Indicator for All Enterprise Adoption

“The standout winner — I guess this one’s pretty obvious — is code. We actually had a thesis that this would be the thing to focus on and that it would go fast, but I think it’s exceeded even our very high expectations. When we started making the models for code back in 2024, our basic perspective was: developers are fast to adopt things, and making models better at code also makes you better at building models. You have this kind of verifiability both during training and inference time — the gains compound, you can build the software to help you build the software. If we just take what’s happening very quickly in coding, that is going to happen more slowly but in basically the same way everywhere else, and we’d have an economy where these models are infused into every aspect of the economy.”

---

The “Country of Geniuses” Enterprise Transformation

“We’re starting to see a preview of what it will look like for enterprises in general, in code. If we were paying ourselves for usage, we’d be one of our largest customers. First we used it just for writing code, but then you see all these things around the process of writing code — managing servers, controlling clusters, visually looking at features. Then you get to the point where, if you can build tools and scaffolding, the model can start to do that as well. The whole thing is picking up steam, and through that it’s just very rapidly changing the way we work. We’ve reached the point where it doubles and triples the rate at which we’re able to produce new things end-to-end.”

Talent & Culture as the Deepest Moat

“I actually think the core element is talent, and the core element behind talent is culture. If you look at our retention rate, it’s the best in the industry and it isn’t close. We’re one of the only companies that has all of its co-founders — we still have all seven. You have to get to almost employee number 20 chronologically to find someone who left, many years after the founding. Last summer, a competitor was trying to hire our researchers by paying numbers ranging from $100 million to $500 million — prices that greatly exceed those of pro athletes. We said to everyone: you’re here for the mission. We’re not going to give you 10 times as much as you’re earning right now just because someone threw a dartboard and your name came up — that’s very divisive and destructive for the culture. At the end of the day, we lost two employees to Meta. The best comparison: OpenAI lost several, and they’re maybe 1.5x as big as us — so even normalized by size, there’s a 10x or 20x difference. We are just clearly doing something different.”

---

Chip & Cloud Ecosystem Strategy — Multi-Cloud, Multi-Chip