Nvidia CEO Jensen Huang talks earnings, China, Trump, and gaming

Indeed, and we are watching these NVIDIA shares affected by what we heard from the company, both on data center and on China. But Jensen, without further ado, I want to bring in Jensen Wang, CEO of NVIDIA. Jensen, it's great to see you. I want to start actually putting China a little bit to the side and asking how you would characterize underlying data center demand ex-China right now. Incredible. Incredible. Obviously, we had a record quarter without China, and we just guided another record quarter without China. And so the fundamentals are this. One, we're in the second generation of our Blackwell architecture. Blackwell is revolutionary. It's got this new NVLink 72 rack scale computer. The entire rack is one computer, not a whole bunch of computer nodes connected together, but it's one connected supercomputer. And it came just in time because of this agentic AI technology that has emerged in the last year or so, which causes the AI to be able to think. It can think, it can do a plan, it could reason, does research, use tools, it could use a web browser, it could use Excel, it could use SAP, it could use ServiceNow, and it could use all of these tools to solve problems, work with other agents, for example. So the amount of computation has gone through the roof. And meanwhile, all of these AI-native companies are coming up with all these exciting new applications. At NVIDIA, we use Cursor. It's our AI programmer. Across the entire company, everybody's using Cursor. The amount of code that we're generating and checking in is just incredible. Our productivity has gone through the roof. And so OpenAI, as you know, had a billion-dollar month. It's incredible growth. If you look at the AI-native companies, last year, they did $2 billion altogether, the first year of AI revenues. The second year of AI revenues, $20 billion, tenfold increase in one year. And that's just the tip of the iceberg. I see AI opportunities, obviously, across all of the clouds and enterprise and industrial applications and robotics and self-driving cars. I mean, just digital biology and drug discovery, we just see that AI is just deploying everywhere. Jensen, you mentioned on the call that China would have been a $50 billion total addressable market this quarter, and then this year, and then you would have seen it go to 50% next year. How do you get to that point with the administration where you can see those kinds of gains? Well, first, I think President Trump really understands and really wants, and it's his policy, to have America win the AI race. And that's really foundational, recognizing that that's important for national security, it's important for economic security, it's important for American leadership around the world. And he understands that the American tech stack should be the world standard, just as the American dollar is the world standard that economies are built on. We want the American tech stack for the world's technology and industries to be built on. And that includes China. And so, if you look at China, China has 50% of the AI researchers in the world. They have lots and lots of AI companies, and they have lots of AI chip companies. Without us there, they will obviously have their own alternatives, and they've been growing. And you've seen some of the AI chip companies have record years, their stock prices are record high. Obviously, there's a lot of opportunities there. If we have an opportunity to go compete for that market, it would be great for America, it would be great for China. And we have an opportunity to go capture some of that $50 billion that is growing at about 50% every year, the second largest AI and computing market in the world. America should be part of that. It's great for exports, it's great for our treasury, it's great for American technology leadership, and it enables us to leverage that leadership all around the world. And so, it's just an incredible decision by President Trump. I'm very, very pleased with it, very grateful for his decision. And we're going to do our best to go capture some of that market. And talk to me about what it means for NVIDIA specifically, Jensen, because, you know, even at full strength, China was about, call it 10 to 15% of the company's sales. And so, I'm curious if you think of China as a nice to have or a need to have for NVIDIA, and whether you need sales into China to keep up the pace of rapid growth that you've been experiencing. Every American technology company, for every American technology company, China is important to have. It's an important market, it's dynamic. The companies there are vibrant, innovative, fast moving. That's one of the fastest growing technology markets in the world. The technologies that are used there, the services that are created there, the open source models, China builds most of the world's open source models today. And DeepSeek, of course, we've spoken about before, really important model. A Q1 from Alibaba, really excellent. Kimi is excellent. Moonshot's model is excellent. There's a whole bunch of really excellent models. Another one just came out, a seed from ByteDance, really excellent. And so, the open source culture is really vibrant there. Those open source models have spread all over the world. They're used by startups all over the world, including here in the United States. It's used in enterprises all over the world. Open source models is making it possible for people to create market-specific, domain-specific, and company-specific models. Really important for American technology companies to be able to run those models really well. And so, I think China's just an important market. Also, as you know, it's just really large. It's the second largest computing market in the world. And I'm just delighted that President Trump sees the wisdom of American leadership, and particularly at this time where AI is taking off. It's vital that America wins the AI race. And so, Jensen, what now needs to happen to get you there, to get not just you there, but other American companies there? How do you overcome that sort of political hurdle? Well, for a while, these American AI technologies are stuck in limbo. And so, now we're getting the engine fired back up again. And so, I'm looking forward to, and I hope that the China market and the Chinese government will allow American technology companies to be able to sell to the Chinese customers. There's a lot of demand there, although there's a lot of alternatives there. The China AI technology companies are really good, as I mentioned. We can be very competitive. You're talking about America's technology industry. It is our national treasure. It is incredibly competitive. And we have every opportunity to succeed in China market if they would allow for American technology companies to be able to play there. And so, I'm hoping that things will get sorted out. I believe there's a lot of demand for H20. The CUDA ecosystem is really vibrant in China. We've been there for a very long time. And so, we can add a lot of value if we can get back in. We've now been approved and licensed to be able to ship to China. Now, we're looking for orders in China, and I'm hoping for it. If it doesn't happen, of course, we've already guided to another record quarter next quarter, and the China business would be a bonus on top of that. But hopefully, we'll be able to address that $50 billion growing very fast AI market long-term. Really important for the United States. Jensen, I want to get a little nerdy here. The networking business, that came in up 98% year over year. You announced a new product where you're able to merge data centers into one super GPU almost. Where does that kind of growth look going into Vera Rubin? You introduced the new NVLink with each new chip. How do you see that then moving with the next, next, next generation? Super nerdy, super good question. As you know, we used to have a network technology called InfiniBand, and we still do. It is still the best scale-out networking technology. In the last five years, we invented three new networking platforms. The first one is called NVLink. It's a scale-up switching architecture. It makes it possible for us to take one GPU, connect across 72 of them in one rack, and turn the entire rack 140 megawatts of power, excuse me, 140 kilowatts of power, 140,000 watts of power into one unified computer. The second thing that we did, that's called NVLink. The second thing that we did was we created Spectrum X Ethernet, a new type of Ethernet, enhanced Ethernet for AI. Not just normal off-the-shelf Ethernet, which is too jittery, too high latency. It's designed for classical networking, not AI computing. We invented Spectrum X networking Ethernet. That allows us to scale out. NVLink scales up. Spectrum X and InfiniBand scales out. And then now, we made it possible for us to connect multiple AI factories into one giant factory. And we call that Spectrum XGS for gigascale Ethernet, which allows us now to scale across AI factories to build these AI super factories. And so, we now have three networking platforms, scale-up, scale-out, and scale-across. And these together, our networking opportunity, you just mentioned, grew incredibly in the last year. Just our Ethernet product line grew to an annualized rate of $10 billion. And that's just gotten started. And so, you could see that networking is going to be a very big part of our business. Now, that's fundamentally recognizing the fusion of computing, networking, and software. That fundamental understanding at NVIDIA for AI almost 15 years ago now, and you know, you guys have been following us for a long time, over the course of the last decade, it's really gone into turbo charge. But this fusion of computing and networking and software is really what makes NVIDIA's AI factories so great. And these factories, as you know, are so expensive, such a large investment. Our performance per watt, our token generation per watt, our computing per watt is, a watt is, you know, energy, computing per power is the world's best. As a result, because your factory is power limited, our revenues, our revenue generation is the best of anybody. And then our performance per dollar, our performance is so good, our performance per dollar is incredibly good, which allows us to deliver the best gross margins to our customers. And then we do this every single year, increasing these step ups, these generational step ups, so that we can help our customers grow, help them drive down cost. And because the token performance is so high, help AI get better and better. Jensen, it's so amazing to see you, put on your technologist hat, but then also put on your sort of diplomat hat, as you've had to navigate all of this different stuff over the past year. Dan's going to get back to another very nerdy question to close this out. But before we do, you and other corporate CEOs here in the US have really been working with the administration to achieve your goals. Of course, the government becoming part owner of Intel, and they've opened the door to other deals. Have you thought about similar deals, other kinds of partnerships with this administration? No, and it hasn't really come up. I think the President Trump and his administration wants to make America great. In all of my conversations with him, that's 100% there. It's about reindustrializing the United States, because we need to have manufacturing here. We need to be able to manufacture critical technology here. We want to have a manufacturing economy here. That vision for reindustrializing the United States is a wonderful vision, and we should get all behind it. NVIDIA is all behind it. The second vision that we're part of is to help the United States. He wants America to win the AI race. It is the most consequential technology of all time. This is the beginning of a new industrial revolution. This technology will be the foundation technology for the next several decades, the next century. It's the most consequential thing that could drive productivity, industrial advance, social advance. He understands that. He wants America to win the AI race. When we speak to him, most of the conversations are really just centered around those two topics, where he ends up at those two topics all the time. He's singularly focused on helping the United States and helping America be great again. I've always enjoyed the conversation with the administration. They're very constructive. They're trying to help. They have a great sense of urgency, which I love. They have a great sense of urgency, which I love. I can't say enough great things about it. Jensen, I want to ask one more question on the nerdy side. Obviously, NVIDIA started out as gaming, still a huge gaming company. You guys build the chip that goes into the Nintendo Switch and Switch 2. The Nintendo Switch 2 is selling like gangbusters at this point, beating out the original sales. There's obviously your own chips. How does the gaming market demand look right now? Obviously, we're going into next year. We're going to see Grand Theft Auto coming out. That's probably going to be on PC at some point. That'll get more people buying more chips. How does that market look at this point for you guys? As you know, NVIDIA's gaming GPUs brought CUDA to the world. It was on the back of GeForce that we brought CUDA to the world and revolutionized computing. It was GeForce that made possible the AI revolution. Now, what's amazing is AI is coming back to revolutionize GeForce for gaming. This last quarter, I think we had a record quarter for gaming between GeForce and Nintendo Switch. As you know, Nintendo Switch 1 is completely revolutionary. I'm really proud of it. It is the highest grossing, highest selling game console in history. It's just utterly unimaginable, the success of Switch 1. Switch 2, off to an enormous start, even more successful, I think, than the Switch 1 started. You can just imagine how far it's going to go. I'm proud of Switch 2. I'm proud of Switch 1. We did several things. We designed a revolutionary SOC, a revolutionary computer chip that delivers incredible ray tracing capability, incredible AI rendering capability called neural rendering. We do it in a little tiny power envelope. It's really amazing that the frame rate of Nintendo Switch 2 is so silky smooth. You don't feel any frames anymore. The technology is really refined, really polished. The images that it's generating is so incredibly beautiful. We use neural rendering. It's amazing. We compute a few pixels. Then we use AI to predict, to generate a whole bunch of pixels. As a result, we save a lot of energy. We produce really beautiful images. I'm just really proud of the gaming business. We're really passionate about it. Very few people ask me about it these days, but I'm really grateful that you do. I'm just incredibly proud of Switch. We're incredibly proud of GeForce. I got to tell you real quick, Cyberpunk looks amazing on it. I think that just knocked people out of their socks because of that. Jensen, we wanted to thank you so much for joining us. Josh, back to you. All right. Our great thanks there to NVIDIA CEO, Jensen Wang. Thanks, Julian. Thanks, Dan. Well, that is a wrap on today's special coverage of NVIDIA's second quarter earnings report. Big thanks to NVIDIA CEO, Jensen Wang. Scan the QR code to get more information on NVIDIA. We'll have more NVIDIA coverage Thursday on Morning Brief at 9 a.m. Eastern time. Have a great night.