Aaron Levie

And I don't think the impact has come anywhere close to peaking yet. And in fact, I would say the rate of progress is exponential right now on at least three key dimensions. So number one is the algorithms themselves. The models are improving at a rate of, I don't know, three to four times a year. They're not just getting faster and better, but qualitatively, they're different.

And I don't think the impact has come anywhere close to peaking yet. And in fact, I would say the rate of progress is exponential right now on at least three key dimensions. So number one is the algorithms themselves. The models are improving at a rate of, I don't know, three to four times a year. They're not just getting faster and better, but qualitatively, they're different.

Remember, we started with pure LLM chatbots. Then we went to reasoning models. And the difference there is with a chatbot, it's like kind of a smart PhD or college student giving you an answer off the top of their heads. The reasoning models, it's more like the PhD saying, okay, let me go off and think about that. Let me do a project on that.

Remember, we started with pure LLM chatbots. Then we went to reasoning models. And the difference there is with a chatbot, it's like kind of a smart PhD or college student giving you an answer off the top of their heads. The reasoning models, it's more like the PhD saying, okay, let me go off and think about that. Let me do a project on that.

And it could work for 30 seconds or a couple of minutes. I mean, as much compute as you want to throw at it. And it will break down your complicated question into a bunch of sub-questions and then it'll try different approaches. And it can validate some of those approaches and come back to you with a much more impressive answer. And if you've been using like,

And it could work for 30 seconds or a couple of minutes. I mean, as much compute as you want to throw at it. And it will break down your complicated question into a bunch of sub-questions and then it'll try different approaches. And it can validate some of those approaches and come back to you with a much more impressive answer. And if you've been using like,

the Grok three deep research or the new chat GPT-03 to do these types of new reasoning models, it's pretty mind blowing what they're capable of. Have we even come close to figuring out how to tap the potential there, especially in an enterprise context? No. But my point is that the rate of progress on the algorithms is again, three to four times. Yeah, you're right. Let me just pull up here.

the Grok three deep research or the new chat GPT-03 to do these types of new reasoning models, it's pretty mind blowing what they're capable of. Have we even come close to figuring out how to tap the potential there, especially in an enterprise context? No. But my point is that the rate of progress on the algorithms is again, three to four times. Yeah, you're right. Let me just pull up here.

Well, I was trying to lay out the dimensions of which progress is proceeding exponentially. So one is the algorithms, which is not just quantitative, it's also qualitative. We didn't even get to the agents part of it yet, but that's the next big leap after reasoning models. We're just starting to scratch the surface there. Then you've got the chips.

Well, I was trying to lay out the dimensions of which progress is proceeding exponentially. So one is the algorithms, which is not just quantitative, it's also qualitative. We didn't even get to the agents part of it yet, but that's the next big leap after reasoning models. We're just starting to scratch the surface there. Then you've got the chips.

I mean, the chips are getting better at, I don't know, three to four X a year. We've gone from the H100 to the H200. Now we're on the GB200. We'll be at GB300 soon. We'll be

I mean, the chips are getting better at, I don't know, three to four X a year. We've gone from the H100 to the H200. Now we're on the GB200. We'll be at GB300 soon. We'll be

No, no, no. They're getting the chips themselves, depending on how you measure it. Each generation of chips is probably three or four times better than the last. Okay. And NVIDIA is back to rolling out new chip, new generation products roughly annually. And I'm just using them as one example. Obviously, there are other companies as well.

No, no, no. They're getting the chips themselves, depending on how you measure it. Each generation of chips is probably three or four times better than the last. Okay. And NVIDIA is back to rolling out new chip, new generation products roughly annually. And I'm just using them as one example. Obviously, there are other companies as well.

So basically the leap from Hopper to Blackwell to Rubin, I guess, will be in next year. And then I think Feynman's coming after that. I mean, really an astounding rate of progress. It's not just the individual chips that are getting better. They're figuring out how to network them together, like with NVL72. It's like a rack system to create much better performance at the data center level.

So basically the leap from Hopper to Blackwell to Rubin, I guess, will be in next year. And then I think Feynman's coming after that. I mean, really an astounding rate of progress. It's not just the individual chips that are getting better. They're figuring out how to network them together, like with NVL72. It's like a rack system to create much better performance at the data center level.

And that would be like the third area where you're seeing basically exponential progress. Just look at the number of GPUs that are being deployed in data centers. So when Elon first started training Grok, I think they had maybe 100,000 GPUs. Colossus was 100,000, correct. Right. Now they're up to 300,000. They're on the way to a million. Same thing with... OpenAI's data centers, Stargate.

And that would be like the third area where you're seeing basically exponential progress. Just look at the number of GPUs that are being deployed in data centers. So when Elon first started training Grok, I think they had maybe 100,000 GPUs. Colossus was 100,000, correct. Right. Now they're up to 300,000. They're on the way to a million. Same thing with... OpenAI's data centers, Stargate.

And within a couple of years, they'll be at, I don't know, 5 million GPUs, 10 million GPUs. And you see that on the power side, right? You're going from 100 megawatt data centers to 300 megawatts. We're just starting to now see the first gigawatt power data centers. I don't even think they're live yet, but this is where they're trying to get to.

And within a couple of years, they'll be at, I don't know, 5 million GPUs, 10 million GPUs. And you see that on the power side, right? You're going from 100 megawatt data centers to 300 megawatts. We're just starting to now see the first gigawatt power data centers. I don't even think they're live yet, but this is where they're trying to get to.