There are now four AI scaling laws — pre-training, post-training, test-time, and agentic — and they form a self-reinforcing loop ultimately limited only by compute.
Jensen outlines four scaling laws in AI and argues they form a virtuous cycle: agentic systems generate data that feeds back into pre-training, all of which is ultimately limited by compute, not by data availability. ✦ AI generated
Jensen Huang · Lex Fridman · 2026-03-23 · original ↗
starts at this moment · 28:15
“So I think you've outlined four of them with pre-training, post-training, test time, and agentic scaling. What do you think, when you think about the future, deep future and the near-term future, what are the blockers that you're most concerned about that keep you up at night that you have to overcome in order to keep scaling?”
I have four scaling laws. As we use the agentic systems, they're gonna create a lot more data, they're gonna create a lot of experiences. Some of it we're gonna say, 'Wow, this is really good. We ought to memorize this.' That data set then comes all the way back to pre-training. We memorize and generalize it. We then refine it and fine-tune it back into post-training. Then we enhance it even more with test time. This loop is gonna go on and on. It kinda comes down to intelligence is gonna scale by one thing, and that's compute.
verbatim transcript · starts at 28:15
28:15We memorize and generalize it. We then refine it and fine-tune it back into post-training. Then we enhance it even more with test time, you know, and the agents, agentic systems, you know, put it out to the industry. And so this loop, this cycle, is gonna go on and on and on. It kinda comes down to basically intelligence is gonna scale by one thing, and that's compute. - But there's a tricky thing there that you have to anticipate and predict, which
28:45is some of these components, it requires different kind of hardware to really do it optimally. So you have to anticipate where the AI innovation's going to lead. For example, a mixture of- - Perfect - ... experts with sparsity. - Perfect. - With hardware, you can't just pivot on a week's notice. You have to anticipate what that's going to look like. It has some- - So good - ... that's so scary and difficult to do, right?
29:09- For example, These AI model architectures are being invented about once every six months. Right? And system architectures and hardware architectures kind of every three years. And so you need to anticipate what likely is going to happen, you know, two, three years from now. And there's a couple ways that you could do that. First of all, we could do research internally ourselves, and that's one of the reasons why we have basic research, we have applied research.
29:40We create our own models. And so we have hands-on life experience right here. This is part of the co-design that I'm talking about. We're also the only AI company in the world that works with literally every AI company in the world. And so to the extent that we can, we try to get a sense of what are the challenges that people are experiencing. - So you're listening to the whispers across the industry, the AI labs.
30:02- That's right. You got to listen and learn from everybody. And have a... And then the last part is to have an architecture that's flexible, that can adapt and move with the wind. And one of the benefits of, of CUDA is that it's, you know, on the one hand, an incredible accelerator. On the other hand, it's really flexible. And so that balance, incredible balance between specialization, otherwise we can't accelerate the CPU,