Live on Bloomberg TV

00:00Daniel Billings with us, Sands Capital Portfolio Manager. So why is the stock down? Well, I think there's a few reasons. One stock had a good run into the quarter and to probably slightly missed sort of the bias and expectations. And it was mainly around, I think, China. And then secondly, black hole, it's still a little bit supply constraints, so ramping it as fast as they can. But I guess the ramp wasn't fast enough. But I would say to us those were pretty strong results. And maybe one interesting fact to look at, if you take out the China revenues completely out of the data center and the rest of the world for them actually accelerated from about 60% growth in Q1 this year to a guided 80% growth in Q3 this year. So actually in video X, China has been accelerating, which which I think is really strong. I think that's an really interesting take. We had Vladimir Galbo from Omdia yesterday talking about how Europe is stop start and certain countries. You've got Denmark really going hard on sovereign AI. Other countries less so. We got told $20 billion for sovereign. Is that strong enough for you, Daniel? Yeah, I think that's a great number. But the truth of the game is still within Hyperscalers and AI startups. I think that's where the real action is for now and it's a much, much, much larger number. And, and I think the interesting aspects there are one, as Jensen noted, that reasoning AI is so much more compute intensive. I mean, he quoted 100 to 1000 times more on the one side and then on the other inside. Those also much better said it reduces hallucinations. And he also mentioned that the return on invested capital you receive from buying these chips is phenomenal, right? He mentioned if you invest 3 million today, you can make 30 millions in token revenues in short order. So I think that's where the real action is as of today. And is that why the real action is talking about where Rubin takes us the cadence of innovation over time video? I think so. I think Ruben will again that the most important thing that India is doing is that increasing the performance per watt. And they've noted that it costs about $50 billion for each one gigawatt data center. Ruben will again improve net performance per watt, which means that Nvidia's competitive situation improves on the one side, but also on the other hand side. You can do more compute, you can build better models with Arup and which means it's a self-fulfilling prophecy a little bit because whatever comes out of these models will have higher IQ and effectively bigger and better apps. You're not worried in any way about Amazon's training, about Google's TPU, the fact that the customers are innovating themselves. Yeah, I know. It's I think it's our job to worry sometimes, but we've looked at this in detail, and I think the idea here is again to say, well, so here's the video. They're building a giant system of a company. They have both networking, CPU, GPUs, hundreds of difference of suppliers, and they put the software on top, they say, and we believe in this, that they are the best performance per watt supplier in the industry by nobody else. And as long as that is true, something like training them cannot compete simply because they don't have the scale, they don't have the system and they don't have the software. So in video to us remains the winner in the sense and that's what's interesting about Jensen Huang trying to say, look, we are more than just hardware in many ways. We are the one stop shop and we're leaning more into that to lean more into the future opportunities Enough for you, the robotics side of things, because automotive remains a bit of an overhang when it comes to China, but thinking also about the quantum and real future here. I think India video is probably the company in the world that runs the fastest that I've ever seen. I mean, what they're trying to do is they're trying to align hundreds of different of suppliers that try to design a GPU on an annual cadence, and they're trying to update their systems on an annual to eight months cadence, which which is unheard of. At the same time, they're focusing on quantum robotics, etc., etc.. So I think Nvidia is running as fast as anything else I have ever seen. And if I may mentioned two, I think Quantum is probably going to slot into the Nvidia ecosystem by saying they have CPU's, they have GPUs and at some point they'll have Quantum somewhere in there too. And you can just switch between those various sort of compute layers. And that was what was interesting with IBM and AMD's partnership, the idea that almost you reference how these reasoning models work. I'm thinking of GPT five, how it just pushes you towards the best possible model for your question. We have the best possible computing standard for your general task. Yes, I fully agree with at the best possible as well as the cheapest and the fastest. And and again, I mean, as we think through the next few years, the question is how much more compute can a video sell? And it's truthfully continues to be a question of the scaling laws that we've seen with Grok for that compute still provides better and bigger models. We're going to have Blackwell cluster scaling in the next 12 months or 6 to 12 months, and hopefully we'll be seeing good results coming out of whatever comes after four and five. And that's what ultimately truly going to drive the best fun video in terms of demand and in terms of the revenue outlook. So you think we should think more about Colossus? We should think more about what matters doing in terms of, well, it's scaling in Louisiana. Should we figure China out entirely considering the government wants to push its own homegrown domestic talent anyhow? Yes, I think we should. And and I would actually argue what China is doing seems to be relatively rational. What China is basically saying is that you're trying to sell that ship that's based off of an architecture that's three and a half years old. And secondly, we've also grown a local architecture or a stock that has caught up somewhat to that old chip. So unless you start selling us a new derivative of black, well, we don't really maybe need you as much anymore as we needed you three years ago. So that all seems to make sense. And I don't know whether the licensing will work out or not. But the interesting piece is if it does, that's a big, big upsell effort for NVIDIA, because Jensen noted it's a $50 billion market and if they're allowed to ship a derivative of the newest chip, their market share, there should be enormous because there's nothing in China that can remotely compete with what Nvidia has.

No schedule data available.

BTV Channel FinderWatch BTV in your area

Sovereign Vs Hyperscaler Demand

Sands Capital Portfolio Manager Daniel Pilling says Nvidia’s growth is likely still to come from hyperscalers. He speaks with Caroline Hyde on “Bloomberg Tech.” (Source: Bloomberg)

More From Bloomberg Technology
All episodes and clips

Bloomberg Markets
"Bloomberg Markets" is focused on bringing you the most important global business and breaking markets news and information as it happens.
More episodes and clips