Ben Gilbert: evaluation

Transcript context

…I'm sure he's in the Navy. He's a great computer scientist pioneer. Yeah, data center. It's big. It's interesting, the objectors to that acquisition—and it's a good objection. This is ultimately, I think, why they abandoned it because they got the regulatory pressure on this. Arm's business is simple. They make the IP so you can license one of two things from them. You can license the instruction set. Even Apple who designs their own chips is licensing the Arm instruction set. In order to use that 20 keywords or so that can get compiled to assembly language to run on whatever the chip is, if you want to use these instructions, you have to license it from Arm, great. If you don't want to be Apple and you don't want to go build your own chips or you don't want to be NVIDIA or whatever but you want to use that instruction set, you can also license these off-the-shelf-chip designs from us. We will never manufacture any of them. But you take one of these two things, you license from us, you have someone like TSMC, make them, great. Now you're a fabless semiconductor company, and they sell to everyone. Of course, the regulatory body is going to step in and be like, wait, so, NVIDIA, you're a fabless chip company, you're a vertically integrated business model, are you going to stop allowing Arm licenses to other people? And NVIDIA goes, oh, no, no, no. Of course, we would never do that. Over time, they might do some stuff like that. But the thing that they were sort of like, which is believable, beating the drum on the strategy was going to be is right now, our whole business strategy is that CUDA and everything built on top of it—our whole software services ecosystem—is just for our hardware. How cool would it be if you could use that stuff on Arm-designed IP, either just using the ISA or also using the actual designs that people license from them? How cool would it be if because we are one company, we are able to make all of that stuff available for Arm chips as well? Plausible, interesting, but no surprise at all that they face too much regulatory pressure to go through with this. Clearly, that idea rattled around in Jensen's head a bunch and in NVIDIA's because—well, let's catch us up to today. They just did GTC at the end of March, the big GPU developer conference that they do every year that they started in 2009 as part of building the whole CUDA ecosystem. It's so freaking impressive now. There are now 3 million registered CUDA developers, 450 separate SDKs and models for CUDA, and they announced 66 zero new ones at this GTC. We talked about the next generation GPU architecture with Hopper and then the grace CPU to go along with it. I could be wrong on this. I think Hopper is going to be the world's first four-nanometer process chip using TSMC's new four-nanometer process, which is amazing. We talked a lot about Omniverse. We're going to talk about Omniverse in a second, but you mentioned this licensing thing. They usually do their investor day, their analyst day, at the same time as GTC. On the analyst day, Jensen gets up there. It's just so funny going through the whole history of this now of looking for a market and trying to find some market of any size. He's like, we are targeting a trillion-dollar market. He's like a startup raising a seed round walking in with a pitch deck.…

Stored transcript either side of the excerpt. The highlighted words are the published quote; the surrounding text is unedited source, never generated.