Well, I'm really excited. We've bumped into each other all over the world at the Paris AI Summit, but this is actually the first time we've gotten to sit down together. So, thank you.
Well, thank you for having me.
So I want to quickly catch people up to both AMD's story and your story. I started covering my first job in tech was covering semiconductors in 1999. Jerry Sanders was still running AMD. It was known at the time as sort of the company always nipping at Intel's heels, barely catching them, small market share. Intel mainly liked it because they could say, 'We're not a monopoly. See, there's AMD.' Obviously, that's nowhere near the position today. Can you talk a little bit about how AMD got to be where it is today? Give people a sense for the scale of it. And also, were there a couple key decisions that you think allowed AMD to flourish while others have struggled?
Yeah, absolutely. So again, thanks for having me. I think you said it best. AMD is a company with a very rich history. We were founded in 1969 as one of the Silicon Valley companies. And frankly, our focus has been on computing all of this time. From the standpoint of how far chips have come, for the longest time, nobody knew what a chip was. Nobody thought about chips. It was something underneath the covers and not necessarily front of mind. For us, it's our business, our history, our heritage. I've been CEO for a little bit over 10 years. We've been focusing for the last decade on building the highest performing computing chips out there. And that is now extremely important. That's why we spend a lot of time here in Washington. We have a global business. Everyone now recognizes that chips are a foundation for national economies and national security. For us, it's about providing all of the infrastructure. When people are super excited about AI, that's great, but you need all the chips underneath it to make sure that you can really turn that AI into something that works. That's really what we do. It's about building the fastest, most power efficient chips that you can. Things are changing. In the past, people used to talk about something called Moore's law in chips, where you would continually double the performance every 18 to 24 months, but it was very linear. Today, it's extraordinarily complex. We're still seeing the same kinds of gains, but it's taking a lot more different tricks because you can't just thin the wiring a little. It's extremely complex. You might call it a little bit black magic, but what it really is is putting together 150 billion transistors in a very small form factor to run all of this AI capability.
So I want to talk, Dario was talking about it on stage. I think a lot of the question is what should America's AI policy look like? And if you could set one or two priorities for Washington, what are the things that would ensure leadership in US technology broadly, but US hardware in particular?
Yeah. Well, I would start by saying I love the fact that we as a country are talking about technology leadership front and center. The fact that the administration released the AI action plan just a couple of months ago, I think that was a very good blueprint. From my perspective, AI is the most transformative technology that I've seen in my career, in our lifetime really. The correct usage or the most accelerated usage of AI will determine the winners and losers both in companies as well as in countries. So for us to be talking about it as a national policy is a great thing. The answer for any technology that is so nascent is that as a country, we have to run fast and even run faster. That includes that we're in a fortunate state that we have an incredibly innovative set of companies in the industry at every layer. We're in the silicon, the chip layer, but Dario was just on the model layer, and there's a lot more in the infrastructure layer. But it is a very competitive world out there. So we need to invest, continue to innovate, and also ensure that we have the right policies. The government helps us have the right policies, including manufacturing in the United States, including making it easier to build data centers, including ensuring that we have enough power to really unlock all of this technology capability.
So there's certainly the aspect of how much do we invest, do we make the right investments in those layers. There's obviously also the how do we negotiate policy with China, and that's been fascinating. We've gone from I remember when the supply chains were almost completely intertwined and people barely thought about it, and now it's obviously a very different story. I'm curious because one of the most puzzling developments in the last year for me has been, no, you can't sell this chip at all in China. Oh, wait, yes, you can as long as you give us 15%. Help me understand because I don't really get it. Is it a national security like Dario saying this is the one thing we have, you're going to ruin it? I haven't heard anyone say, and I've heard other people say we should just sell and if we don't make them, they'll make their own and better they use ours. This was the first time I've heard anyone say there's a percentage that if you give the US government somehow it makes it good. Help me understand how this fits into policy.
Yeah, sure. So let's take a step back and just look at what the landscape is. I would start with, look, there's no question national security should be our number one objective. That has to be first and foremost above all, and we all believe that. From a technology standpoint, the fact that we are leaders in AI is actually a great thing for the country, and we need to continue to not only maintain that leadership but really extend that leadership. Then you go to the next level and you say to yourself, okay, the world is a global place. People need technology. We have a lot of capability, and we want to make sure that the world is also innovating on our AI stack. What that means is there's no one company that has the answer for all of the innovation in the world, or any one country frankly. The innovation in technology comes through a lot of learning across different ecosystems, and we want the world innovating on the US AI technology stack.
So I take it you disagree with Dario that we should just not sell to China because chips are the only thing?
I would disagree. Yes. I would disagree. I think there are ways to really think about how we protect national security as our number one priority, which we are. Our most advanced chips are export controlled, and they should be export controlled. But there is also an opportunity for us to get an AI stack that is based on American technology out into the world, and I think that's a good thing. That's really what we have been focused on, which is finding that right balance.
I want to shift gears a bit and talk about this infrastructure boom that we're having, because it seems like it's not only growing like this, every month it seems like we've gone from billions to trillions, this huge scale. Wall Street's super excited, but I'm also curious, when Oracle is a big customer of yours, when they announce they're going to pour hundreds of millions, billions, hundreds of billions into these data centers, does that mean you guys are getting tens of billions, hundreds of billions? How big a deal is it for you that Oracle, which was kind of the company you turn to if you needed a little extra cloud capacity, but not one of the hyperscalers? What does it mean for you that they're building on such big scale?
