Bring in show music, please. Hi, I'm CNBC producer Katie Kramer.
Today on Squawk Pod, Lisa Su, CEO of chipmaker AMD, on the promise of artificial intelligence for her $380 billion company. The signals in the market right now say that we're still at the very beginning of the technology, but you're starting to see significant productivity. And for how we live, where we work, what we buy, the chips we need to power AI. I don't think it's a big gamble. I think it's the right gamble.
And the flying tomato, Olympic gold medalist Shawn White on the new opportunity in snowboarding. Where's the next Shawn White? I'm like, they're there already. We're just not telling their story. From the halfpipe to big business. It's been deemed action sports for the longest time, but it is sports. Plus, the rest of today's news that got us squawking, like the Trump administration reportedly weighing curbs on powerful proxy firms and index fund managers. Most people do not have the ability, time, resources, energy, or effort to look through those votes.
It is Wednesday, November 12th, 2025. Squawk Pod begins right now. Stand back by in three, two, one. Cue it, please.
Good morning everybody. Welcome to Squawkbox right here on CNBC. We're live from the NASDAQ market site in Times Square. I'm Becky Quick along with Joe Kernan and Andrew Ross Sorcin.
Good morning. Thanks all here. Meanwhile, the government and airlines warning that disruptions have cropped up during the shutdown could end up continuing even after things reopen in DC. The FAA now instituting flight cuts in order to keep things safe. As you know, at US airports during a press conference at Chicago's O'Hare airport, transportation secretary Sean Duffy saying that easing those restrictions will depend on air traffic controllers coming back to work. And he warned of further chaos if the government doesn't reopen. Big disruption thus far, massively more disruption as we come into the weekend if the government doesn't open. So the house has to do its work.
Flightware is saying that about 1,200 flights have been canceled so far. That's just today. Yesterday there were about 1,900 cancellations. Airlines are supposed to be cutting about 6% of their schedules.
Reports say finance industry executives have been invited to dinner with President Trump at the White House tonight. JP Morgan CEO Jamie Dimon, NASDAQ CEO Adena Friedman are said to have been invited. Among the other potential guests reported by Bloomberg, Goldman Sachs CEO David Solomon, BlackRock CEO Larry Fink, and Morgan Stanley CEO Ted Pick.
The Trump administration reportedly now looking at steps it could take to reduce the influence of proxy advisory firms like ISS and Glass Lewis and index fund managers including Vanguard and BlackRock. The Wall Street Journal reporting that one option could be broad shareholder recommendations. Another could target recommendations on companies that have hired the proxy firms to do consulting work. That's a conflict that we've talked about on this broadcast that other CEOs have talked about as well. Additionally, the report says the Trump administration is looking at forcing index fund managers to vote on proposals in the way their clients vote.
I think that's a little more problematic. My own opinion there. Proxy firms have faced heavy criticism from corporate leaders including Elon Musk who called them terrorists. Jamie Dimon has made some comments over the years just about the influence and some of the conflicts. One of them being this idea that you know if you're playing on both sides. The idea though, I don't know where you guys land on this. To me, the idea of either preventing them from making recommendations, there is a First Amendment. You should be able to do that. I have no problem with that. I think getting rid of some of the conflict part of this is important, but also the idea of taking away the vote from a BlackRock or a Vanguard or somebody else and just saying it's up to the individuals who are trading or own the shares.
So do I. It's interesting that you say that. The default what happens then I think could be wrong. They just go with the company. Most people unfortunately don't vote at all. So what you are effectively doing by default is giving, you talk about who's got the power. You're giving more, this is a trade-off. Who do you want to have the power? Do you want the management to have the power or do you want ostensibly Vanguard and these other funds should be representing the interests of the shareholders? That is at least ostensibly what they should be doing, which is different to some degree than what the managers necessarily are doing. It's not always a line.
I just don't like that. I don't like giving them my power.
No. But just to be clear, Vanguard, BlackRock, all of these funds, if you want to vote yourself, you have every right to vote yourself. You can go on and vote now. Almost everybody, most people don't. Almost everybody's giving you the right to it. It's not that they're taking your vote from you, but most people don't. It's just that most people are not.
I just don't want them speaking on my behalf.
Well then, if you're saying then you don't want anyone to speak on your behalf and you don't want to vote, then what you're saying is you want the management.
But I don't want them to. If I choose not to vote, that doesn't mean I'm giving you my vote. The late great Charlie Munger even said who made Larry Fink king. There was a time when at a shareholder meeting somewhere he said we need to force behavior. This happened to be about climate change and whatever you believe about climate change. But he was voting for activists that were forcing fossil fuel companies to do stuff based on Larry's own personal woke ideas, which he's totally backed off.
I'm hearing you. My question to you though is a different one. If you're saying that you don't want to vote.
No, I'm not saying I don't want to vote. There are times when I do vote. By the way, I have money in all of these places, too. There are times when I do vote. But if I choose not to vote, it doesn't mean that I'm giving you my vote.
I think then there's a question. I would think that you should be able to give the vote to the firm or not, and that should be an option. But it shouldn't be an opt out. It should be an opt in if I choose to give you my vote. The question is whether you think that most Americans whose 401k plan is effectively being managed by these funds, and the reason you're having a target-date fund, which I would never invest in, but there's a lot of people that have target-date funds or have S&P funds in their 401k plans or other things that are being managed, and the reason you're using them is these are the firms that you're hiring to do this.
No, these are the firms my company are hiring. So these are my choices in my 401k.
And the question is whether they give you, if you get the choice, you can use the choice, and if you don't use the choice, whether the firm should be able to have that choice.
But it should be an opt-in, not an opt out. And if I decide, I like this company. I want to own it. I like what management's doing. I'm long their stock. What's the problem with defaulting to doing where the company has more influence than some outside proxy firm that might have all these ridiculous notions in their head?
I just think the conflict. Let's get rid of the conflict piece. Are you talking about CEO compensation that you don't want that rubber stamped, or are the corporations always doing something bad?
No, the question is who is arguably going to have more of your interests? The management. If they're aligned with the shares going up like you are, then they'll have. But the question is, who's looking out for you? Is it going to be the management or would you like to have somebody else do that?
You put your money into that stock for a reason because you like the management, you like what they're doing.
So here's one issue I want to lay out so everyone's on the same page. The reason why people don't vote and the reason why this is an issue is most people have money in an S&P 500 fund, and so there's 500 companies in that fund. Each of those companies has a proxy vote every year that could have a dozen or 20 different elections that have to be made. Most people do not have the ability, time, resources, energy, or effort to look through those votes. So then the question is if you're saying that you just blindly say whatever management says is great, go with God, that's one view.
But it's not whatever the proxy owners say. I would defer to them before your friends in the proxy.
I'm not talking about the proxy guys. Ostensibly what you're trying to do is the reason people hire people to manage their money is because they are spending money and resources to try to look at each individual issue arguably for you.
Not index funds. The reason I invest in the S&P 500 is because I don't want to pay somebody a lot of money to manage that stuff. I certainly don't look there. There may be times when I would agree with the proxy firms on some of these things. There would be times I would agree with management, but if I'm not taking your vote away, I'm not blindly giving it to them either. If I opt into that, fine. But the default should not be that they get my vote if I don't take it.
And so how would you like that to work? Because what's going to happen under your scenario is people will have a 401k plan when they're first enrolled in the plan 25 years ago, they're going to have a check box that says would you like me to do it or would you not like me to do it? Some people will just check it blindly, some people won't. With opt-in, you're going to get a much smaller percentage of people who are doing that.
Even if they never vote again and never touch it, but it's like who do you put your blind trust in? Do you put your blind faith and trust in the head of the proxy of the manager? Your default is it seems to be that the corporation's trying to put something over on you.
If you had a contract of any sort, I assume you would have a lawyer look at it for you. The reason you'd have the lawyer look at it is because ostensibly they work for you and are trying to look out for your best interest. The people that you're buying the product from hopefully do. When I buy a pair of sneakers, I hope that they have my interest. You see an index fund manager in that light. I don't at all. Put the stocks in there, make the fees low, and I'll take my chances. I'll handle it. Don't be voting your own ridiculous notions with my stuff. You're not king. There was a no kings day. Don't forget that.
You recognize who you're making the king then. The CEO.
I trust that's why I'm buying their stock.
But you're buying I'm buying American business.
Next on Squawk Pod, Lisa Su, CEO of Chipmaker AMD, had a big message for Wall Street and consumers. AI is for everyone. The full conversation with Lisa Su is right after this break.
Welcome back to Squawk Pod. I'm Andrew Ross Sorcin along with Joe Kernan and Becky Quick on a Wednesday morning. The end of the shutdown could be close. The House expected to vote on the continuing resolution later today to reopen the government. You're looking at a beautiful live shot right now of the capital as the sun comes up. Let's hope a deal gets made against that backdrop.
It's almost like it comes up every day no matter what.
Thank God. The sun will come up also rises and the sun will come up tomorrow.
AMD hosting an analyst day yesterday. It's the first time in three years. The chipmaker says that revenue growth will expand 35% a year over the next three to five years. And the company's AI data center is expected to grow at 80% a year and potentially hit tens of billions of dollars in sales by 2027. Joining us right now is the AMD CEO, Lisa Su. A note on this, AMD stock up phenomenally. You can see since she took over, 7,589%. Lisa, I think when you took over this was a $2 billion market cap. I think today it's above $386 billion. Congratulations.
Thank you. Who's counting? Well, it's great to be here with you guys this morning. Thanks for having me.
It's great to see you, too. Last week you had earnings that came out that beat all expectations and yet the stock actually declined. I think people were waiting to see what you said at the investor day. They liked what they heard, especially when it came to the revenue, the annualized growth that you're expecting. 35%. I think most people were looking for closer to 30%. Predictions are hard though, especially about the future. And I guess the big question is what makes you so confident?
Well, first of all, it was a great day yesterday. It's always nice to sometimes take a step back from just what's happening on a quarterly basis to really talk about long-term strategy. And it has been, as you said, three years since we've hosted an analyst day. But there's been so much that's changed in the market over the last 12, 18, 24 months that we needed to take a step back and talk about the company as a whole. Frankly, we're in just an amazing cycle for technology and specifically around AI. The market is accelerating at a pace that we just did not understand until over the last few years. We started a couple of years ago thinking the AI accelerator TAM might be $300 billion, and that sounded like a really big number. Each year we took it up, and now when we take a step back, we see this massive opportunity, really over a trillion dollar overall market as we get through 2030. It's a place where everybody needs technology. You can think about computing as equating to intelligence. And if you can get more intelligence, why wouldn't you want more computing? That's the market we're in. We're really excited about what we have. It's a product portfolio, a set of partnerships, but it's also a vision for where computing is going in the future.
What makes you so sure of that demand? What do you see that makes you think not only that this demand is going to be big, but that it's real and that it's not going to disappear?
Well, Becky, we spend a lot of time in the ecosystem. Being in this industry for over 30 years, you see a lot of cycles and a lot of technologies take off. The key is what people are saying as they're actually utilizing the technology. We work with all of the largest hyperscalers in the world, all of the largest enterprises. If you looked at the AI trajectory over the last couple of years, people started with tremendous excitement. ChatGPT came on board and everyone was like, wow, this is pretty good. Then enterprises started using it, and the early enterprise usage was a little bit slow. People were trying to figure out what to do. But what we've seen in the last 12 months is every one of our largest customers saying, we can see the inflection point now. Demand is accelerating because people are now starting to get real productivity out of the AI use cases. All of the largest hyperscalers in the world are investing more in CapEx because they can see the return on the other side of it. We're always looking for the signals in the market, and the signals in the market right now say that we're still at the very beginning of the technology, but you're starting to see significant productivity improvements.
Do you really think that they can see the returns on the other side, or do you think there's a little bit of a big gamble going on? A bunch of them will tell you quite openly, look, I'd much prefer to be ahead than behind. I'm worried that if I don't invest enough now, I won't have enough capacity later, but I can't necessarily fully math out how it's going to work out on the other end either.
I don't think it's a big gamble. I think it's the right gamble. The way I would say it is we can definitely see that more computing and more investments equate to a faster pace of innovation. I think that's definitely there. You can also see all of the pointers to more productivity and more capability. Exactly timing of ROI, whether it's 12 months or 24 months, that's where it becomes a little fuzzier. But our largest customers are extremely disciplined companies and they have the balance sheets to invest.
The one that everybody talks about is OpenAI. OpenAI in 2028 is expecting to lose $74 billion. By the way, they have $13 billion in revenue right now. They believe that somehow between 2028 and 2030 they will all of a sudden become a profitable business. So they have to make a swing from a $74 billion loss. Something magical has to happen in the course of 24 months. So the question that I would ask you as somebody who has to think of them to some degree as a customer or creditor of sorts, how do you think about that math?
I would take a step back and say when you look at the entirety of the AI technology companies, you have the largest hyperscalers, and you have what I call the AI native leaders like OpenAI, Anthropic, and others. The underlying foundation is the same; they all need computing. Relative to your comment about OpenAI, I think OpenAI is a phenomenal company. Let's start with the vision, and then we'll come to the math. The vision says you need more computing. Today, computing is a limiting factor in how much people can adopt AI. You have to invest ahead of the curve so that you can get the entire user base on there. If you take a look at the growth in users, not just at OpenAI but across the entire AI ecosystem, we're seeing that growth because there is true productivity coming at the other end of it. We are all disciplined business people, but we see an incredible opportunity in front of us. If one has the capability to address that with a fantastic balance sheet, you would absolutely do that.
Let's talk about your plans for AI data centers. You are scaling pretty rapidly. You see the growth there in terms of revenue being greater than 80% over the next several years and potentially getting to have more than a 10% market share in that business over the next few years. To do that, you're going to have to go through Nvidia because by some estimates they have more than 90% of the market share right now. What's your plan for doing that?
This is an amazing market with tremendous growth. Whenever you have a market that is early in the cycle and on a steep growth curve, it really favors innovation and partnership. From our standpoint, technology roadmap, we're super excited about the new products we're bringing out. Our MI350 series is currently in production. We have a big step up with new rack scale architectures coming next year. We also have a really strong set of partnerships with OpenAI, Oracle, Meta, and all of the large hyperscalers. The key in this type of market is that everybody wants to have multiple solutions. There's no one solution that fits all. With different workloads, you're going to want the best technology for each one. We do believe we will have a significant piece of this market. Between large hyperscalers, enterprises, and sovereign nation states building large data center installations, we are an excellent option.
Mindboggling what I've seen with AMD. We spoke briefly about foundry, Jerry Sanders, one of the co-founders, who I knew for a while. He started this in 1969 at Fairchild. Every employee got stock options when he started. He's still alive, maybe he might be watching. He's 89 years old. The reason I bring this up, Lisa, I watched AMD challenge Intel for years, and I have to rethink everything because I thought of it as an also-ran. It always seemed to be copying or cloning a lot of things that Intel did. The chart doesn't even work for Intel versus AMD over 10 years. Something happened somewhere. Can you tell me where Intel missed something or AMD capitalized? What caused Intel to lose its preeminence?
First of all, Joe, we love Jerry. I hope Jerry's watching. He absolutely founded a fantastic company, and I'm honored to be CEO. I've been CEO for a little over 11 years. What you find in technology is when you're at inflection points, you have to make the right strategic bets. You have to decide where it's going to go and invest in those places. That's what we've done. We had a vision that high performance computing was going to be the most important thing, and that's what we invested in. With that, we systematically gained confidence. To answer Becky's earlier question, when I started, we had 1% of the CPU market, and today we're probably about 40%. I think we have a path to be over 50% or number one in the market. That came from a true belief of where the technology is going and how we partner with the largest companies out there.
I am not anybody to question your ability to do it, but it is really hard to come up with this rack scale infrastructure. You care about putting your customers first and building trust with them. You're not going to do anything that would put that at risk. But even Nvidia stumbled when they brought out the Blackwell chips, their third iteration. So what are you doing? How can you be so confident that this pulls off without a hitch?
I can say that our track record is we've built an execution machine, and we are very confident in our ability to bring out new products. Over the past 5 years, we've invested over a hundred billion dollars, both organically and inorganically in the company. Organically, we built up our engineering capability, hardware, software, and systems. Inorganically, we've done $60 billion of acquisitions to bring in the talent required to pay off these big bets. The key in technology is you have to make the right decisions, and I think we've made many right decisions. We recently acquired a company called ZT Systems that specializes in data center rack scale solutions, and they have been absolutely key to our roadmap going forward, along with a number of other hardware and software acquisitions over the last few years.
Lisa, big question in the industry is how to think about the depreciating value of these chips. You're talking about this fantastic new technology and making it better every single year. As many people are building these data centers, the fundamental question is how long do these chips actually work or are meaningful if you have to upgrade them given how quickly things are moving every two or three or four years, and what happens to them afterwards?
What we've actually seen is, I think this was something people were wondering: is the next chip so much better that you have to rip out your current gear and put in new gear? That's just not the case. As AI workloads expand, think about it this way. The most advanced foundational models need the latest and greatest and will take the newest chip as soon as it comes out. Then you have a whole bunch of other AI companies servicing regular demand, medium-size models, maybe even small models. Large enterprises may do quite a bit of fine-tuning on their models. They don't need the latest and greatest. What they want is something very reliable and steady. They'll use the current technology. What we're finding is there is a large set of use cases. We only always talk about the bleeding edge, but what I'd like people to think about is AI is for everyone. Everyone should get a piece of AI. That means in the cloud, you may use the current generation technology. By the way, I'd love you to be using AI in your PCs because I think that will also be one of the big opportunities.
I totally get it. The big question is, are we building railroad tracks that are going to be here for a very long time? Are you going to have to replace the tracks every 5 years, 10 years? This phone is the iPhone 17. I just found an old iPhone 2. You can't use the iPhone 2. You can't really use the iPhone 10. You probably maybe the iPhone 12. This is the fundamental question in terms of all the money going into this space and how those chips get used. I agree with you, they'll get used in some way somehow. But this goes to the math of all of this.
I think the math of people who are thinking four years is too long are now beginning to see you can definitely use these assets for five or six years, and some would say even beyond, for more of the workhorse applications. That's what makes the growth rates what it is. For the absolute bleeding edge, new technology, you're going to build big new data centers with different form factors. But the rest of the assets will continue to be used. You need a 5-year depreciation schedule for that to really pay off.
Going back to your earlier point, just looking at some of the deals from the outside, it looks like a crazy whirlwind where anybody and everybody in the industry, both technology and energy, are meeting all the time. Mike Worth was here talking about how he's talking to somebody about building out an AI data center in Texas. Is it like that from the inside? Is it this crazy whirlwind where you've got to meet with everybody all the time?
It is a whirlwind. As we look across the ecosystem, one thing has changed. As a silicon company, we used to talk to our customers and our customers' customers. We would talk to the hyperscalers and end users like the enterprises. I think what's becoming clear is now we need the entire value chain, including the people building the data centers, providing the power, and providing the financing. To your point, all of these things are important for us to put together. It's a broad ecosystem. The way I think about it is that's the way it should be because we need all of the pieces to accelerate to put on all of this compute as fast as possible.
