AMD CEO Lisa Su goes one-on-one with Jim Cramer

Lisa Su
2 followers
3 Video ViewsยทJan 13, 2025
AMD CEO Lisa Su joins 'Mad Money' host Jim Cramer to talk the latest tech advancements, competition in the space, the state of AI and more.
After a roller coaster of a year, AMD, one of our favorite chip makers, one owned by my
travel trust, basically fly for 2024. And remember, AMD is the only chip maker that
comes close to NVIDIA in this artificial intelligence space. Yet now the stock's
down roughly 80 points from its highs this spring. Something's wrong here.
So let's check in with Dr. Lisa Su, the chair and CEO of Advanced Micro Devices,
get a better sense of what's going on. Dr. Su, welcome back to Myanmar.
Jim, it's wonderful to be here with you.
Well, Lisa, I saw you get honored earlier this week. It was a terrific event, Time 100,
and you were a featured toaster because you've done so much well. And you said something really
caught my ear and attention. You said that we'll look back 10 years from now and we can't believe
what life was like now. Will you tell me what you're seeing 10 years from now that would make
it so this seems all so pedestrian? Well, first of all, Jim, you know,
great to be here with you. And I have to say, I really do believe, you know, we talk a lot about
AI, but we're just really at the beginning of understanding what AI can do for us. And,
you know, if you take a step back, like I've been in this industry for like over 30 years,
and we've seen, you know, a lot of interesting technology. I mean, we saw cloud, we saw mobile,
we saw PCs, we saw these were all great technologies. But AI is something different
because it makes everything better in terms of, you know, it makes our businesses, you know,
more productive. It makes our engineers more productive. It makes researchers more productive.
And it helps us really, you know, I can see a world where AI really helps us solve, you know,
really some of the world's toughest problems. And that's why I say, you know, it was a nice event
on a Monday night. It was one of those events where you could see, you know, a decade from now,
nobody's going to be asking, what was the ROI on AI? They're going to say that AI is part of our
life. It's how we live. It's how we run our businesses. It's how we do our research. It's
how we teach our kids. That's the world that I see. What do you say to people right now who are
watching who just say, you know what, that's just businesses talking to businesses. I don't see it
impacting my life at all. Oh, I beg to disagree. I really believe that AI will impact everyone's
lives. It's just starting today. So, you know, let's not be impatient, right? Tech trends are
meant to play out over years, not over months. We've only been in this, let's call it, you know,
chat GPT world for maybe like 18 months. We're all learning. It's fun. You know, we all use it.
We use it to make ourselves, you know, to answer questions for us. We use it to help us
in sales and marketing. You know, we use it to help us build chips, frankly, in terms of large
language models. But now we're going to see it in all aspects of our lives. And, you know, I think
of it as, you know, personal assistance. I think of it as education. Probably one of the areas that
I'm most excited about is, you know, what we can do in health care and what we can do in
accelerating research. And these are things that, you know, if you look at the process that it would
take for, you know, drug companies to discover new drugs, it might take like, you know, years.
And most fail, 90% fail.
Exactly. If you could change years to weeks or days, that's kind of the superpower. And that's
just one example. And, you know, the beauty of all this is you need the computing. And, you know,
that's what we do, right? We do all of the foundation of the computing that enables, you know,
AI to help us in all of these industries. And, you know, that's why I'm just so excited about
where we are right now.
All right. Now, I want to disabuse some people of a notion. People tell me, well, listen, it's either
NVIDIA or it's AMD. And to me, if you don't think that there's room for two, you really don't
understand the space. But you taught me that. Let's teach others that.
Well, it's absolutely true. You know, the way to think about it is there's no one size fits all
in computing. There's no, you know, there's no, you know, only one architecture. Actually,
you're going to need the right compute for each application. Like, I'm excited about AI
everywhere. I'm excited about AI in the cloud. I'm excited about AI in PCs, you know,
so that we each have our own, you know, sort of AI engines. I'm excited about AI at, you know,
the edge and infrastructure that we use, you know, throughout, you know, the entire world. So
there are all kinds of different computing that you need. And the beauty of AMD is that, you know,
we've had this vision for the last, you know, 10 plus years. This is kind of how we've thought
about, you know, so what do we want to be when we grow up? We want to be the high performance
computing leader. And AI is a huge, huge piece. OK, but let me give you how I feel as someone
who's followed you from when the stock was at five and you explained to me why to sell Intel at 50
and why buy AMD at five. So far, so good. But you explained to me a roadmap and it was a roadmap
of Italian cities and it was going to lead to triumph. And when that triumph occurred, your
company would be worth a lot more. I don't hear much about the roadmap anymore. I just hear much
more about AI. What happened to the roadmap? Is that not making you that much money anymore?
Well, first of all, let me take a step back and say you need computing of all sizes. So if you
want to talk about Italian cities, I could talk to you about Italian cities. There's really,
think about it this way, Jim. There's like, you know, three types of computing, right? There's,
you know, this big accelerator AI computing that we're talking about for our large language models
and for, you know, all of that stuff. You know, we're at Dreamforce. MI300. Yeah, MI300. We're
at Dreamforce this week. There's a lot of talk about what you need to do, you know, with agents
and things like that. There's traditional compute and in CPU land, where our Italian roadmap has
done extraordinarily well. If you think about, you know, data centers around the world, they need to
become more efficient. You know, there's too much old equipment. That equipment has to be upgraded.
You need the highest performance, lowest power efficiency, you know, overall total cost of
ownership. You know, we have, we're now in our fifth generation of our Italian cities. So Turin
is our latest and greatest, you know, products there. We're just about to launch it in the next
quarter. Lots of exciting things going on there. I talked about AI PCs that are also a very exciting
trend that we're going to see going forward in the PC market where we have, you know, intelligence
in AI. And then, you know, we also have our embedded computing, Jim, you know, I've talked
a lot about that, you know, sort of the, what we're doing with FPGAs and adaptive computing,
you know, aerospace and defense. We use a tremendous amount of AMD underneath that.
When you think about industrial test and measurement. So we are computing across the
board. Obviously there's a lot of excitement around AI and accelerators, but they're really,
you know, you know, three major legs of our computing stool. Well, it makes sense that the
AI part is exciting because you put out some pretty amazing numbers for 2027. A lot of people
initially felt that they were impossible as we go on. It seems like that they're reasonable and
rational projections, aren't they? Yeah, absolutely. And we see the computing market in, you
know, AI accelerators getting to $400 billion by 2027. Why do we don't care about the Italian
roadmaps? Hey, we love all compute. We love all compute. But when you think about the power of AI
and where it goes, you know, sure. That's where a lot of our focus is. You know, we've made
tremendous progress in, you know, our roadmap around accelerators. You know, MI 300 is our
current generation. We're now on a one year cadence. So every year we come out with, you know,
new hardware such that we can make AI even more efficient and more capable going forward. So,
we're, you know, we're launching our next generation here again in the fourth quarter.
Then you'll see us with another new generation, MI 350 series in 2025 and then MI 400 series in
2026. So our Italian cities plus our accelerators, you'll see, you know, just a tremendous growth
opportunity. And if we saw you in the same room as Jensen Wang, would you be able to say, you know
what, I get along with the guys, decent competitors or are things just too heated between the two
companies? I'm a big believer in the technology ecosystem really works when you have, you know,
good competition and frankly, you know, good partnership as well. And there are places where,
you know, we partner across the ecosystem and then there are places where we compete. But the
main thing is customers want choice, right? Our customers are the largest customers in the world.
They want the best tech, you know, everybody wants to have the good choices and, you know,
that's what we're offering. We're offering just a very, very, you know, strong, you know,
high performance computing roadmap. And that world works when demand exceeds supply, doesn't it?
It is one of those cases, Jim, where, you know, people have asked me over and over again, you know,
are we going to have enough? And we are building, we are building, you know, a ton so that we have
enough computing capacity, but demand is very strong as well. And so, you know, we're going
to continue to ramp up the supply chain, but I view this as, again, let's not think about it as,
you know, what happens, you know, over the very short term. Let's go back to how we started with,
like, we're going to look back at this whole time, you know, 10 years from now,
and we're going to say, you know, what was life like before AI?
I'd say I look back 10 years from right now, and I remember where your stock was. I remember you
told me what I was going to do, and had I said buy it at five and sell it at seven, well, I guess we
would have missed a lot of points. I want to thank Lisa Su, AMD chair and CEO. Don't grade it on a
day when the Fed gives you a 50 basis point cut. That's not how to look at stocks, businesses,
or Lisa Su, Mad Money's back after the break. Coming up, is Workday rising? The other
enterprise software star jumping for generative AI, next. Don't miss a second of Mad Money. Follow
at Jim Cramer on X. Have a question? Tweet Cramer, hashtag Mad Mentions. Send Jim an email to
or give us a call at 1-800-743-CNBC. Miss something? Head to madmoney.cnbc.com.
