Lisa Su in Conversation with Ryan Patel | SXSW 2024

Lisa Su
2 followers
4 Video Views·Jan 10, 2025
AI is future of computing and the single most transformational technology innovation of the past 50 years. In this opening keynote session, Dr. Lisa Su, Chair and CEO of AMD, joins host Ryan Patel to discuss the future of AI, how millions of people use AMD-powered services everyday, and AMD’s vision to solve the world’s toughest challenges with leading-edge high-performance and adaptive computing.
How y'all doing? Nope. My first time here at South By. I didn't even peek out of the thing. I need
more. Come on. How y'all doing? If you're going to stay with me, I promise you a couple things.
This is not going to be an ordinary keynote session, fire set conversation. You know why?
Why? Because you got stuck with me. That's why. And I'm so fortunate to have AMD to allow this.
We're talking about future of AI. I want to give you what I see on the future of AI in my lens.
So you know how we're going to do that? Not just words, but since we're in Austin, with vibes.
I'm going to show you some videos that I saw that I felt that was important. I have a special,
you know, depending how your energy is with me, I may have a special guest or so,
maybe. And some new looks of something. But I want you to know technology isn't just about words.
It's about the product. It's about the feel. It's I want you to feel this. So why I said this
conversation is going to be different. It's going to be a little difficult, too, because we're going
to try to show you and talk about things that I saw that was interesting. So so I've got the
pleasure to welcoming the legend in my eyes, Dr. Lisa Su. So I told you, I'm going to give you,
I'm going to give you what you want. My promise, our promise. You brought me here, Lisa. I did.
Is to have you walk away with a couple of takeaways, give you what you want to see. And
right now, you know what I'm going to give you exactly what you want. They want pictures of Dr.
Lisa Su. I'm going to give it. Welcome, Lisa. Thank you. This is like a great crowd. Is this
a great crowd? Wonderful. Welcome to Austin. Thank you for coming. Yeah, I told Lisa that we
were going to sit down immediately that we came up. And apparently I didn't do a good job of doing
that just either. But, you know, we're going to flow. Yes, we are. Yes, we are. So I appreciate
you all. We appreciate you. We're going to start this conversation. So typically, you know, when
you know, as you can probably tell, I'm pretty comfortable with Lisa. We've met before a few
times, a few times. And she when I think of and I got to meet a handful of you all from the audience,
actually, you know what? I'm going to switch something up. See what happens when you give me
energy. I change things. You know, I was talking to a couple of people yesterday when I was walking
around the Expo floor about what AMD is and how it has evolved. You've been there for 10 years and
they were telling me about you, actually. Oh, no. What did they say? Well, it sparked something that
you told me a while ago that when you started, you run to problems 10 years ago when you took
home as a CEO. When I say to run to problems, what does that like? What does that stand out with you
right now in your philosophy? Well, first of all, I have to say I am absolutely thrilled to be here
this afternoon at South by Southwest. It's an incredible honor. And I've been like super lucky
throughout my career. You know, I'm an engineer by training. So semiconductor device physics, if you
really want to know. And I've been in the industry for 30 plus years. And, you know,
when I used to tell people that I was in semiconductors, they would be like, what?
Like, what's that? And I would say chips. And they're like, chips? You mean like potato chips?
And I'm like, I know that's what you were thinking. Don't even start. No, no, no. I mean,
semiconductor chips. OK, you need semiconductors to power everything that you do, whether you're
talking about the largest supercomputers or you're talking about, you know, PCs or your car
or what you have in your house. All of that is powered by semiconductors. And what is so cool
today is everybody knows what semiconductors are. I think we've all realized that semiconductors are
essential. And that allows me to be here to talk to this great audience at South by Southwest.
And yeah, I believe in running towards problems because the idea is, you know, we're all like
working like night and day on things. You might as well work night and day on something that's
really important. And that's what I'm so fortunate to do, is work on things that, you know, help,
you know, shape the world and make the world a better place. So. Well, I think why I showed you
the video was it's not millions of people use it. It's billions of people that's run by on your
service. Is that right? That is right. You know, I venture to say everybody in this audience goes
through an AMD chip a few times a day. And, you know, if you think about it, if you're if you're
using Microsoft Teams or Office or you're on Zoom or you're on any of the Google services,
you're probably running through AMD somewhere in your infrastructure. If you're a gamer,
how many of you are gamers in the audience? Some gamers. Yeah. Wait, wait, wait. Come on.
You know, you know what? They're still sleeping from last. OK. All right. So if you're gamers,
you know, Microsoft, Xbox, Sony, PlayStation five, Steam Deck. Well, your portfolio has changed.
Right. And that's the that's the to me, that's interesting, because we were talking about the
chips. Right. You mentioned a couple of those names. That's right. That's right. We have just
lots of stuff. We're in Tesla. So if you're a Tesla fan or Subaru fan, you see AMD everywhere
in the ecosystem. And that's what we want. We want to make your experiences better. And that's what
we do. I mean, real quickly, you should talk about your background. Graduate from MIT. She's
one of the few that I know, like I know I know you knew you love semiconductors from the very,
very beginning in college. Am I wrong? Yes, that's true. That's true. I was a nerd at heart. Don't
say that. Yes. Don't say that. I was. So I went to MIT for my undergrad. And, you know, what do
you need to do when you're in college? You need to find a job. And so my very first job was freshman
year. It was a I would say it was a grunt job for a graduate student in a semiconductor lab.
But I got to make these really tiny little chips that were the size of a dime or a quarter.
And I was just amazed at everything that you could do with them. So, yeah. So I was in
semiconductors when it wasn't sexy. And I would say I don't know that it's sexy now, but it's sexier.
No, no, no. So I would say it is because last night I was with your James Knight. I was watching
an Oscar viewing party at the Soho House. And I'm sitting there pretending like I won one.
And so why I was there was because you all have done a lot of stuff in Hollywood. Right. We tell
my avatar to a couple other movies. Yeah. Look, I mean, it's it's so great to be working with the
best in the industry. And that's what we enjoy. So when you think about Hollywood, when you think
about making, you know, these, you know, these really feature films, when you think of all the
special effects that are required, what they require is a tremendous amount of compute.
And, you know, we were very proud to be sort of really associated with some of these studios.
And the idea was, how do we give as much compute as possible so that the creators can really
accelerate their rendering, accelerate all the production that needs to be done?
And, you know, some of the things that we've done are like, you know, it used to be that,
you know, the processors, if you wanted to know the processors that were in these
machines that were used in creative studios, maybe had like, you know, eight to ten cores or
something like that. That's that's as technical as I'm going to get. And we said, look, we're going
to change the game. And we put a tremendous amount of computing power in there. So you could really
take what used to take, you know, days to render could now be rendered in hours. And that just
changes the way you can make movies. So, you know, we're super honored to be associated with some of
the best, you know, Pixar, you know, Elemental was built on, you know, AMD, AMD hardware, you know,
what effects just amazing, amazing capabilities. If you think about Avatar 2 that was on AMD.
So we've had just a planet. We just saw Planet of the Apes. Yes. Yes. You know, we saw that
war is over last night, won an Oscar. That's what I was celebrating. War is over. If you don't know,
won an Oscar yesterday. And so we have a very special guest. We decided to bring a special
guest for the South by Southwest audience. David Connelly is here with us. So shall we bring him
out? No, not yet. From what I ask, why? You know why? I got to briefly talk to him in the back. We
didn't even know he was going to show up because he showed it left four third in the morning
yesterday because he was at the Oscar party. Not better to come on stage after that first look.
Then David Connelly, executive VFX production for Vetta. Welcome. Thank you.
How's it feel? You big, big winner. Great. I feel awesome. First of all, it's so exciting to be in
front of so many enthusiastic people. I'm very excited to be here. Fresh off of some exciting
news from last night. We won an Oscar for an animated short called War is Over. Songs inspired
by John Lennon and Yoko Lennon. And we did it entirely on AMD using real time technologies
to animate something that has never been done before. And somebody.
And I don't you know, when I think of AI, I think of technology, I think of industries that are
changing. David, you wanted to be here when I talked to you last week. Yes, absolutely. And
you thought I was much cooler than Elton John and those folks, right? Absolutely. He's not here for
you. He's here for me. Lisa made him come in. We've got you got to be here. We got to be here.
No, I we we appreciate this because I think you were at the at the cross sections of what the
future looks like. And so out of those out of that reel that you just saw, is there anything else that
you want to add to give it to the audience? So a lot of the work that you saw, there's a mixture
of legacy filmmaking and a lot of new technologies. But one of the things that I want to just share
very quickly, I met Lisa in January of 2019. And we started talking about the intersection of media
entertainment, technology, and what it means to to create art and what it means to create movies.
And there is an intrinsic link, as Lisa stated in her opening statement between compute power,
technology and art this today. And so we wanted to really leverage the work that AMD is doing to
help us create images that are more fantastic than we could have thought that we could do 30
years ago. 30 years ago, Peter Jackson started a company, one computer, eight people, and did some
work on a movie called Heavenly Creatures. Now we have over 1,800 people and hundreds and thousands
of cores and computers and machines all over the world creating movies like Avatar 2. And you saw
some sneak peek footage there of Kingdom of the Apes. I had to ask. I told you we're going to give
you something. Lisa wanted to come with something that we wanted to share. And so we were able to
do this. So I appreciate you. Yeah. So the thing that I want everybody to understand is that these
images are possible because there's a partnership with a company like AMD that gives us the leverage
and the insights. And they work with our teams and we work with their teams to help optimize.
Because a lot of this wouldn't be possible. We would still be rendering Avatar today if it wasn't
for a lot of the partners here. I know some of you know what I mean. I heard it because I heard
the chuckle come. Oh, I see the head nod. I see you. So anyway, thank you very much, audience. This is so
great to be here. We love the partnership, David. I can't thank you enough. I think the key
in our partnership is sort of the two-way feedback that we have. And so, you know, we really appreciate
you. David, tell us a little bit about what's next. I mean, you're using every piece of
compute we give you, which we love. But a little bit about AI. And how do you think AI is going to
change what you do? So first of all, we use every bit of AMD. I mean, every bit. As soon as it comes
off the assembly line, we're at the door knocking, can we use it? Secondly, I do want to
talk about AI. And I'll talk a little bit about the future. But I want to acknowledge one thing.
The use of AI in the entertainment industry is still a sensitive subject. I think a lot of people
misunderstand it and don't really truly understand where AI can help us as a community. But I also
want to make sure that artists understand that AI is a tool and that this is not about replacing
artists. Where I see things and where I'm hoping to go is that we, as an industry, have moved from GPU
to CPUs. And there's an intrinsic relationship between the types of technology that AMD is
working on. I'm believing that the future of filmmaking is a partnership between passive
entertainment and active entertainment. And this is where we're going to start seeing the
intersection of games. I know I heard a loud roar from all of you gamers out here. I want to hear
another one. Because gaming, visual effects, and moviemaking is all the same. And I want to jump
from video games to movies. I want to go from active entertainment to passive entertainment.
And I want to go into real-time entertainment. And that is where films are going to be in three
to five years. And we're not going to be able to do that without the help of companies like AMD,
where we get real-time processing and real-time rendering. And a lot of AI is going to help us
get to that place. So I appreciate that. Because I think what I'm hearing from both of you,
that the creators out there and those who are artists, that you get to spend more time
on the actual art. So I expect both of you all, that means it's only going to get better,
right? The production, the films, the visual effects. That's right. So part of what we need
to do is there's obviously a lot of pressure to create content and create content in shorter
deadlines. So part of the business of a duress deadline delivery industry, like ours, is that
we want to maximize the amount of time artists spend at their stations, dreaming up fantastic
things that none of us can. Because I always look for the young kids, the older people,
everybody to come together and create something fantastical. I don't want them to hit a button
that says render and wait for four hours, or four days, or in some cases, four weeks.
That doesn't help any of us, because you as artists, you feed off the creative energies.
Well, I appreciate you. I actually tried to get the Oscar, so we can take a picture,
but they told me I couldn't do that. Can't have that. I tried, but I appreciate you coming on,
David. I really appreciate you, man. Thanks for coming and flying in for us. Lisa. Thank you so
much. Thank you very much. I mean, it's amazing, right? I mean, to be able to talk about this,
and you think about Hollywood, and I also think about creators. Where's the creators in the room?
So, on this side. You got more followers. I would imagine there's a few more. Creators?
Yes, on this side. We'll take it, Lisa. All right. But when I think about Hollywood,
we think about the creators that I hear more from this side. I think about the things that
you are working on. You've got a partnership with Adobe. I remember talking to Shantanu,
and they were so excited, because it is all about that. Can you talk a little bit more about that?
Yeah, absolutely. I mean, if you think about it, our goal in life is to get the best compute out
there, whether you're a Hollywood producer, or you're a content creator, or even if you're just
like an amateur, you like playing around with videos. We give you the most compute and the
lowest power. We love to work with Adobe, frankly, because they're always pushing the envelope,
and the idea is, again, for us, when we optimize hardware and software together,
we get a better product, and you're able to get sort of what you want to render out much faster.
Now, generative AI is the thing. I mean, I know we're going to talk a bit more about AI, but
when you think about what generative AI can bring to the whole content creation process,
this is the opportunity that we have to take computing to the next level. And as David said,
this isn't about replacing creators. This is really about making creators much more productive,
so that you're able to do in far less time, you know, much, much higher quality capability,
and that's what we do with our hardware, as well as our software partnerships.
Well, let's stay there for a second, because now I want to be here for, you know,
we're here at the heart of AI. What do you think the evolution of hardware and software, right,
you know, is there's going to be a need to go a lot further, right? And where are we at today,
for those that understand it, and where we're going to go in the future?
Yeah, so I think the key with, you know, bringing AI, and really now let's think about AI over the
next 10 years, is that you need different types of hardware engines. So we'll talk about, you know,
CPUs and GPUs, and you might even hear the term NPUs. The idea of all of these are trying to get
more compute in as small a footprint as possible, so it's cheaper, faster, better, so that you can
get more done. So when you think, if I said AI, and you had a crystal ball,
Lisa's crystal ball, I think I like that tagline. Where do you think the future is,
as of right now, in AI? Where are you making bets? Well, I think the best thing to kind of keep in
mind is, I think AI is the most important technology that has come, you know, sort of
on the scene, you know, over at least the last 50 years. I mean, people talk about it as, you know,
really comparing it to the Industrial Revolution. So as much as the internet was important, as much
as, you know, PCs are important, mobile phones are important, you know, AI, and especially
generative AI, has become the most important thing. And there's a good reason for that,
frankly. The good reason for that is, you know, if you think about it, you know, ChatGPT came on
board, I don't know, 14 months ago, something like that, and it just captured everyone's
imagination. Like, AI has always been around, but the ability to make AI so simple that you can just
say, hey, you know, I want to know what I should do in Austin, Texas, this weekend, for example.
Having that possibility only comes with a tremendous amount of computing power.
So I do have some show-and-tell. Can I bring some show-and-tell? I haven't seen this yet,
so I'm excited. So if you think about, you know, I said chips were like the size of a
dime or a quarter, but that was like, you know, 30 years ago. They're kind of a little bit bigger now.
So what this is, is it's actually our latest generation, a generative AI chip. It's called
MI300, and it actually has 153 billion transistors. It has 12 little chiplets. You'll see the
little chiplets, and we actually stack them, you know, side by side, sort of in a two and a half D,
as well as on top of each other to get the most power in one of these chips. And why is this
important? When you think about ChatGPT and how it works, it actually works with thousands of
these, or tens of thousands of these, in massive data centers that are there to really, you know,
both train, and then when you ask it a question, it needs to go back to these data centers to do
that. So it just gives you a sense of the power of technology, and you know, I love stuff like this.
Did you want to look at it? If you don't mind. They don't trust me. They're very valuable,
Why are you laughing? I'm going to drop it. Geez, you're supposed to be on my side, too. Like,
I'm right here. This is pretty cool, though. I mean, I mean, one of the unique, one of its kind.
No, no, no, there are thousands and tens of thousands. But more to come.
But look, so the idea is you need big AI. So this is what I call infrastructure AI.
But you actually, if you look over the next 10 years, you're going to see AI everywhere. You're
going to see it in the biggest data centers in the world. You're going to see it at the edge.
You're actually even going to see it in PCs, you know, maybe on something like what you're
carrying there. By the way, I got to use this. Hey, Brittany, I'm keeping this laptop. It's got
the AMD Ryzen chip in here. I wanted one, and now I'm not giving it back. It's yours.
You know, when you said something also, AI, I believe a buzzword, if many of you haven't heard
it yet, for you all is AI PC. Who's heard of AI PC before? You've heard that word? Two people.
They're just not admitting they've heard it. Yeah. Well, if you haven't heard it,
you're going to be hearing it over and over again. And I want to spend some time here on
why I think many of why this is important, because that is the evolution of peace in the PC game. So
when I say AI PC, since there was only two people who raised their hand, what does that mean? What
does that mean? How fast is it coming? Yeah, so absolutely. So AI PCs, you know, I talked about
the data centers and going into this big iron infrastructure that sits somewhere in the cloud.
But when you think about AI PCs, it's actually the idea that everyone, everybody in this room
can have your own AI capability. And, you know, I'm sure that you've noticed, you know, if there
are lots of people on chat GPT, when you ask it a question, it might actually take a while.
And now if you're thinking about going from, you know, text to image or text to video,
again, that's relying on a lot of stuff out there in the cloud.
You know, the goal of AI PCs is to make sure that every one of us has our own AI capability.
You don't have to go and go out into the cloud. You can actually operate on your own data. You
can actually ask it questions. It'll answer for you. It'll answer it for you faster. It'll answer
it for you in a private manner, because maybe you don't want your data going everywhere. And
it's just the beginning of what I think is the ability to make all of us much, much more
productive, whether you're just doing, you know, sort of your regular email, like you want to ask
your, you know, when is the last time I saw Ryan Patel and what did we talk about? That's something
that you can ask your AI PC, or you can actually ask it to create, you know, new things for you if
you're a creator. So I actually have another little example. Maybe I can show that, Ryan,
if that's all right. This is an example of our latest Ryzen 8000. We call it our, you know,
HawkPoint system. And we thought we would have a little bit of fun with it, because some of you
may have heard of this thing called stable diffusion. What it does is it allows you to,
you know, write something and say, you know, write something in text, and it'll make it an image.
We decided to have fun with Austin. So if you guys have seen Congress Avenue, you might have
also heard that we have bats in Austin, and you might have also heard that, you know, tacos are
good in Austin. So we decided to render an image. We asked it to render an image on an AI PC of a
very cute bat having tacos in Austin. They're going to just run it one more time so you can just see
how fast it is in terms of rendering. This is on a local PC, okay? You don't have to pay anybody
anything to run this. And we were missing the salsa to begin with, so we asked it to add salsa.
Oh my goodness. At the end. When I had tacos yesterday at South Congress, I didn't see them.
You didn't see the bats. Fortunately. You know, bat season is, it's probably in a few weeks,
so you might want to come back. You know, some of that. You know, how, you know, when you were
saying this, I think for the audience too, like, we're close, I mean, we're close to AI PC. Like,
people, you think. We're here. We're here. Think about, you know, AI PCs. You can go buy an AI PC
today. Like, for example, we just rendered on a HawkPoint. You can buy these systems today. But
more importantly, this technology is just going to get so much better. And it's not about just,
you know, what we do. It's about what we do together with our partners, you know, to unlock
this capability. So we have a little bit more than half time to go, but I'm going to tell you,
put questions and comments. I will read them. We're going to take some comments,
meaning if it's something funny, but put it in there. I'll read it. If it's good enough,
it gets upvoted. I'll look at it. So put your questions in now that we have it.
You know, when we think about, what do you, when you think of AI, right? I want to stay there for
just one more second, right? What excite, you know, you talk about your excitement. What are
things that we need to do as an ecosystem, right? Because I think you have been one of the few,
like, I've talked to other leaders in your space with your partners, and it comes back to you,
that you bring in, no matter who it is, together to make sure that we,
as a community, are moving toward the right spots. Well, first of all, when you look at AI,
it is, it does take a village, you know, frankly, to put these things together. We love our
partnerships. We love our partnership with Microsoft, for example. You know, Microsoft
has been leading the way with, you know, Copilot, and, you know, that's running in Azure in the
cloud. We're partnering very deeply on AI PCs as well. We're spending a lot of time with some of
our OEM partners, like, you know, HP and Lenovo, to really put together the entire ecosystem there.
And then it's about the software partners as well. So our goal is to make this super easy
for all of you guys to use. And, you know, that's the promise of AI PCs, but more importantly,
that's the promise of, you know, AI, you know, going forward.
Now, see, I have this question, and I haven't heard the response to this,
I've seen in your interviews just yet. And we're in Austin, AMD's headquartered here in Austin. So
this is almost like, for me, seeing her, it's like a homecoming. I love the boots. If you've
not seen the boots, they're pretty, pretty cool. Yeah. What are you doing AI internally, right?
We talk about this stuff, can you share a couple tips, things that you do that maybe others can
kind of go and maybe take it back to their organization? Yeah, absolutely. Now, small
correction. AMD is actually headquartered in Silicon Valley. You're based here. It's headquartered in
Silicon Valley, but we consider Austin the unofficial headquarters. I live in Austin,
I love living in Austin. That's what I meant. So the, look, you know, one of the things about AI
that I like to say is, you know, people are worried about, you know, AI is going to replace
people's jobs and stuff like that. That's not the way I think about it. I think companies that
learn how to leverage AI are going to win over companies that are not leveraging AI. So at AMD,
we want to be at the bleeding edge. We're using AI through every aspect of our business. We're
using it to design chips, we're using it to design faster chips, make them more reliable,
build better software, make sure that we're able to, you know, really, it's a productivity tool.
And my goal is, I'm telling my engineering team, I'd like to, you know, increase the number of
products we can get out every year with the use of AI. We're using it, you know, throughout our
HR processes, our financial processes, our customer service processes. And again, this is a way of,
let's call it moving up the food chain, because we're allowing our team to get some of the,
you know, less fun things done by AI so that we can really get, you know, higher value added for
our employees and, you know, really have it be more fun. Well, I wanted to share that because
for many people, you know, who are leaders in their companies are just, they don't know where
to start, right? Even to have this conversation, you feel like everyone has to have a kind of an
AI strategy or just an AI ability just to be able to talk about it, right? I think it's an AI
strategy. I think the other thing to remember is it's not like it's perfect, okay? AI is not perfect.
It is, we're all learning along the way. I would say that I'm personally learning. So,
every day I'm learning new things about what the technology can do and how do we need to shape
our entire ecosystem and our work in our next generation products as well. So, that's what
makes it fun. Well, you told me something when I saw you in September that you're using it for your
organizational personal account. If you don't mind, share that. So, I'm absolutely using Copilot
and it's one of the things that we use when you think about, you know, very easy things to do,
you know, summarize meetings, track action items, you know, make sure that we're, it doesn't write
my emails that well, I have to say. Is that why I got to respond? No, no, I don't use it for that,
but look, we're all experimenting with AI for sure. And then we talked about software a little
bit, but, you know, I also think that there can be some obstacles, you know, when you are in these
partnerships in there, like, what do you think about the software? What are some things on the,
when you think about the industry, the proprietary approach, you know, also it is very competitive
in what you do. Yeah, so certainly for right now, people are trying to get product to market as fast
as possible. I mean, that's the thing about AI. This industry is changing at a pace that I've
never seen before in terms of just how fast things are going. And as a result, you know, you want to
be as productive as possible. There are various ways to do that. You know, AMD's approach, our
approach is to have an open ecosystem. We are huge, huge supporters of the open source. We believe that,
you know, there's no one company who has the answer to everything. We actually have to collaborate.
So, you know, things like, you know, PyTorch, you know, that's an, you know, excellent example of
how we can make machine learning and AI programmers much more efficient, because you can program at
one level, and you can use sort of any hardware systems, you know, out there. Things like Hugging
Face that have a set of open source models as well. These are very important partners for us.
You know, OpenAI's Triton, you know, JAX. These are all examples of a software ecosystem that is
open, that's meant to really give people the power of, once you, you know, write software at some
level, you can actually use in multiple hardware environments. The fact that she, I mean, how do
you stay up with all this stuff? You personally, like, you are, I mean, well, I'm going to answer
the question I'm going to ask her, because I know she loves product. Building new products, you are,
she's so passionate about it when you see her speak about it. Is that how you keep up with all
these things that's going on? Well, I spend a lot of time with the engineers, but frankly, I mean,
we're living in a time where everyone's learning every day. So, I've never learned so much as
what I've learned over the last year in terms of where technology is going. And that does help,
you know, in our world, you know, for our next generation hardware systems, we're actually making
big bets, right? So, the things that we're working on today, you're going to see, you know, three or
four years from now. And so, we have to kind of decide, you know, what's next, but that's our job.
What's next? Well, that's, I was going to say that's a secret, but no, it's not a secret.
Look, you're going to see AI everywhere. I mean, that's, you know, we are constantly trying to make
it faster, easier to use, more accessible. I'd like to see everybody have an AI PC. You know,
that's not going to happen this year, but there's this whole opportunity over, it'll start this year
over the next couple of years. I think you'll see it throughout product portfolios. I'd like to see
everyone have access to all the knowledge that you can get, you know, from AI that comes from
making our chips more capable, you know, over time and deeper partnerships. So, yeah, those are
the types of things. And that's what I wanted you to hear. I wanted you to hear that this change is
actually not going to take actually that long. It's coming and everyone's going to have access
to it, as you mentioned. I didn't really understand it until last year that this was here and that we
will be in this time next year, but this time next year, you're all at South by, many people will have
this capability of having it and it's only going to get more. So, the change is coming and hopefully
you got that sense and that it is here. I also want to thank, so your 10-year anniversaries
of this, of being at the helm is coming up, right? That's right, that's right. I'm just about into my
10 years as CEO of AMD. So, you think about, you know what, I'm glad that you all clap because
that's not an easy task. You think about, well, hold on, you think about that. Are you trying to get rid of me or something?
I was trying to keep you, I'm trying to keep you here, right? You got 4 billion annual revenue
to more than 20 billion last year and I can't, I mean, to be that long and to continue to push,
that's why I called you a legend. I know you're so humble that she won't admit it to that degree,
but like there isn't people, there's levels to this and she's got that and so for me, I would
love to hear some of your, we're going to take some questions now, but leadership strategies, like
I want to dive into you as Lisa Su, who's done this, there isn't, there's only one,
there's only one and I'm just, I just hope that people can see what you've done and I want them
to under hear kind of some of your insights and tips when you think about leadership strategy.
Yeah, well, look, I have to be honest, I mean, being CEO of AMD was absolutely my dream job, so if you can
say that you're doing your dream job, that is actually what I'm doing and what I really
enjoy about it is you get to really work on products that matter and so when I first took
over as CEO, like any company that doesn't have quite enough resources, we were probably
doing too many things and it was very important for us to decide what we were going to be good at
and what we're good at is building the highest performing computing capability out
there, so we're a high performance processing company, that's what we spend our time on and it
turns out that that was not a bad bet because frankly everyone needs, you know, high performance
computing these days. I mean, not a bad bet, I mean, people laugh and know what I'm talking about,
the market likes you and you, rightfully so, what's one word to describe, okay, I'm going to put you
in a little bit of a fast, not fast fire, but I just thought about something you said, how would
you describe in one word your leadership style? I think I'm a doer, a doer, I like to do things.
I know you are, that's why, that's the, that's the, that's the, I, I, most, you know, most leaders
wouldn't say that, that they're a doer. Well, you know, my, my view in this world is we get lucky
enough, you know, first of all, I'm so thankful for the team that I have, the, you know, we have
25,000 people, they're fantastic and our goal, my goal is to set very ambitious goals
and then help the team get it done, so that's why I call myself a doer. I love it, I know you are.
I love going into the labs, I love seeing what's the latest and greatest going on and I love
spending time with our customers and partners. I called her, I asked her, I'm going to come, when I
come to Austin, go see the office and I did and she allowed that to happen, but while you were
just talking about your employees and hopefully, I think the crowd will allow me for a minute to,
I know you, you didn't ask me to do this or you didn't ask me that this, this opportunity, but
you know, I posted something, AMD is streaming this, live streaming for all their employees
globally. If you had a message for them watching, hopefully it's okay, if you had a message for them
right now, what would it be? Well, first of all, I was walking out of my office at noon and there
was a thing of popcorn that was at the front door and I walked out and I said,
what's the popcorn for? And they said, it's for you, because they were all getting popcorn to
watch this keynote. So look, I am so blessed to have the team that we have and as we say,
we're only as good as our last product, so let's get the next one done. That's good, but where's
my popcorn? You could have brought popcorn, we could have been here together watching it together
live. So I got some questions up here that I'm seeing. Let's carry, I'm going to read it.
Thanks, Garrett. I'm going to give you a shout out. Curious to understand how semiconductor
technology intersects with energy use and our climate crisis. What topics in the space are
a priority focus for you? That's a great question. Yeah, Garrett, that's a great, great question. So
it turns out that to build more powerful machines, more powerful compute, you actually use a ton of
power and we're actually at a place where there's an intersection of not just, you know, can we get
more performance, but frankly, can we cool these systems? I mean, these systems are massive, they
require a whole bunch of energy. Our goal in life is to build the most energy efficient systems out
there and that is a huge focus for us. We have, you know, pledges to reduce, you know, sort of the
amount of power that we take in our computing technology by, you know, 30x over the next five
years and it's a big focus for us. Man, there's a lot of questions coming up here. I don't know
how I'm going to get through all of it. There's some AI people in the audience for sure.
What percentage of AI interference do you anticipate will happen on the edge versus in the
cloud? I think it's Stefan. Great. Yeah, look, I think, so just for the audience, you know, when
you think about inference, inference is basically when you ask a question, you get a response and a
lot of that, almost all of that happens in the cloud today and frankly, that's some of the latency
that I talked about. It takes a few seconds for it to answer. I do think a significant percentage
is going to happen at the edge, whether in clients or in places closer, you know, closer to your
person and it's just going to be a matter of, you're going to see AI in all of these different
places. We do think that, you know, inference will surpass training as we go through the next,
you know, year or two. So that's where we see the AI trends going. This next question, I'm
going to move just a little bit to the right. Let me see. What is Lisa's workout regime? We want to know.
That's a little personal. You got stuck with me.
I told you I'm giving you what you want. I would say that I maybe walked out a little bit slower
today. I have a great person who comes to my house every day or every few days anyways and
it's boxing and strength training. That's why I keep distance, keep the distance through it. I'm
just telling you. She doesn't share that by the way. I knew that. You told somebody to put
it in the prompter though. It said anonymous, so yeah, I did actually. No, I didn't. Let's see what's
up. You want to, that next one's good? Sure. All of them are good. Yeah. How are you, I mean, there's
so many. Thank you guys. I really appreciate y'all. We're here for y'all. How are you dealing
with NVIDIA's dominance? Can AMD catch up? And I'm rooting for you. I'm not saying that. It says
that right there. Philip, I see you. We can answer anything. Well, look, what I would say is, you know,
AI is really the most important technology over the next 10 plus years. We're so early in the
current cycle. I think you're going to find that there's no one-size-fits-all for tech. There's
never a one-size-fits-all. We feel, you know, we have a lot of respect for NVIDIA, a lot of respect.
They've done tremendously in there from a roadmap standpoint, but we also have a lot of confidence
in where we're going. I think, you know, we actually launched our MI 300 line back a few
months ago, and we had tremendous support throughout the ecosystem. And you'll continue to see us
with AI in every aspect of our products going forward. Love it. Here's another one from Megan.
I manage AWS's, is that you, Megan? Hey, Megan, how you doing? See, I told you we're taking questions.
So Megan's question is, I manage the AWS global demand plan for GPU products. One of our big
bottlenecks is chip supply. Lisa, do you see ways to increase future supply and unblock this?
Absolutely. I can help you out, Megan, I promise. He's smiling, by the way, in the front. Big grin.
Yes, but look, you know, in the chip industry, we do go through these cycles where, you know,
demand is higher than supply. One thing I can say for sure is we're really good at ramping up supply.
So you will see, you know, the supply for GPUs increase significantly every quarter, you know,
going forward. And that's our job. Our job is to build capacity. There are factories being built
all over the world to increase supply going forward. And it is true, I've now been with a
number of, you know, research institutes and companies where they say that they're being
bottlenecked because they don't have enough compute. And every time I hear that, I'm like,
that's what I'm working on. I'm working on making sure that we have enough compute
to really unlock all this creativity that's out there.
Great question. This is a comment, but I have to read it because I'm just,
I just have to do it. Sorry. Sorry, Lisa. Dr. Lisa, I have a summer internship interview with
AMD on Thursday to be an AI application developer. You're already funny. Let me finish.
Can you recommend me? Did I put you on the spot? I didn't mean to do that.
No, no, no. Hey, look, we want the best and brightest at AMD. So all of you who want summer
internships, we're always open for AI app developers and my team will definitely be
reaching out. So thank you. Yeah, that's what I'm talking about.
Very nice. Hey, you know what? I like it. I like your confidence. That's the, you got to do it.
You got, that's, I love it. I told you I would read your comments. I said I was.
Okay. Wait, I have now one because, you know, we think about the start of internships. What is
one thing that you in your career that you think you've, you've accomplished that you're most proud
of? Wow. Let's see. So probably the thing that I'm most proud of, and there's so many, you know,
sort of good, uh, you know, good products and great stories that I have to work on, but I'll
tell you one, uh, which was, you know, I became CEO in 2014 in, in about 2015 or so, you know,
my CTO, you know, Mark Papermaster, uh, who's been my partner, um, all of these years came to
me and said, you know, Lisa, you know, I think we can compete to build the most powerful
supercomputer in the world. Uh, and now you have to understand we hadn't really, you know, gotten
our news and product line or our new Epic processors out yet. And so that was a bold thing for Mark to
say. Um, he said, look, it's going to take us, you know, at least five years and we're going to
work on it. And I said, sure. Okay. Uh, great. Great. Let's, uh, I w I'm not going to put it
the revenue plan just yet, but please, uh, please work on it. Um, I'm, I'm very happy to say that
today's world's fastest supercomputer is, um, sitting at Oak Ridge National Labs. Um, and it's
called Frontier and it's powered by AMD, uh, AMD, uh, Epic CPUs and, um, and GPUs. And, uh, you
know, one of my most fun things is to represent the company. And so for the opening of, uh, the,
uh, the supercomputing center, it's actually one exa flop of supercomputing. I got to go
see the computer and sign our name on behalf of AMD. So that was, that was a very proud moment
for me. So that's, that's a great story. Um, here's another one question. AMD gives a lot of
influence. Um, I'm going to read Niels. Thanks for the question. I'll maybe reword it a little bit.
AMD has a lot of influence or has going to have more influence. Influence comes with responsibility.
In your opinion, what is AMD's responsibility when it comes to AI?
Yeah, so for sure. I think all of us as, you know, large companies in the AI ecosystem have
a responsibility. Uh, we actually have a responsible AI team at AMD. It's actually led by, um, our
president, uh, you know, Victor Pang. Um, I think the idea is to make sure that, you know, as we
bring out hardware, that it is in fact secure, um, as we work in the ecosystem, um, ensuring that,
you know, when you think about the things about AI that we have to be careful about, it's how you
train models, making sure that the data that you're using isn't biased in some way. It's not giving
you, you know, right now, I think there's all this conversation about, you know, are the, are the
models giving you the right answer? Uh, I think what we want to do is we want to make sure that
the models are giving you the best answer they can. And so we do have a responsibility in, in
how we implement, um, you know, sort of the data that goes in and some of the models that are there
and, you know, how we use the data, uh, making sure that we're very, uh, protective of people's
privacy, um, and, um, intellectual properties. So those are absolutely, uh, things that we take
very seriously. But what I will say though, is the answer of those who are a little bit, um, worried
about what AI will do, it's not, the answer is not go slower. Okay. That is definitely not the
answer. The answer is to be very cognizant of some of the risks. And so, you know, we think about
that, but, uh, this technology is so powerful and so transformational. We must go faster. We must
experiment and just do it with a watchful eye. Um, yes, I, I see you, I see you working. Um, real quick
before we get, uh, I want to answer this last question. How is AMD pushing the limit of Moore's
law from a chip design and manufacturing standpoint? Maximilian, I know you know your stuff. That's a
good question. Yeah. So for everyone, Moore's law is the idea that you can, um, you know, increase
performance, reduce the cost of semiconductors every two years. Uh, it's really governed the
industry for the last 30 or 40 years. Um, it is slowing down, uh, without a doubt Moore's law is,
uh, slowing down. But you know, one thing that I found, uh, super interesting in, uh, you know,
sort of the, the, the world of technology is, um, engineers are super smart when you give them a
problem. And so when you say that, Hey, Moore's law is slowing down. Okay. That just means that
you're not going to get everything from traditional device physics and scaling. Um, what you have to
do is, um, come up with new things. And so for example, we're stacking chips. Um, like I said,
we're stacking chips side by side and on top of each other, uh, we're putting a much more, you
know, design capability and design software, uh, design, um, hardware, software interactions,
and all of those are meant to extend, uh, you know, Moore's law and extend, um, sort of the
performance curves that we have. So this is my favorite portion. She kind of knows it's coming,
but she doesn't really know the question. So it's called Ryan is curious, which means I'm just
looking for one word or a sentence. You ready? Okay. Okay. I may even take one from, from the
audience too, from there. Um, F1 or golf? Golf. F1's okay. She's a great golfer, by the way. I'm
What does a weekend look like for you outside of work?
I like doing pretty much nothing. This is why we get along. That's okay. But actually for,
for those of you who are visiting Austin, um, you know, my husband and I will often, you know,
go out on a Saturday night. There's some great restaurants in Austin. So hopefully you've gotten
a chance to experience some of them. You have a favorite cuisine. Uh, right now it's Japanese.
You know, I, I love Soto. I love Uchi. I love, you know, the right now it's Japanese.
Um, I'm going to use this. I'm going to use this one as a Ryan's curious. Thanks. I think it's
Ty who asked if you could give one piece of advice. So I'm going to ask you to do it in
a sentence if you don't mind. Um, if you can give one piece of advice to Lisa in her twenties,
who will be living in the AI era, what would it be? That sounds like a job. That's a job question
I think. So can you do that in a sentence? I would say, and by the way, this is the advice
I give to people. I would say, um, you know, all education is about learning to think it's
about learning to problem solve and whether that was 30 years ago now or 30 years from now, I think
it's the same. I think, you know, the technology changes over time, but if you can learn how to
solve problems, you're going to do great things. So I'm going to interrupt this because I think
we've got some people who are aspire entrepreneurs in the room who want to make a difference in,
in, in the space. Any other tips that you think if not just giving yourself advice, but what,
what would you, where should they start? What should they be looking at right now?
Well, I, you know, I think there are so many interesting things, you know, whether you're,
you're, you have a hardware background or a software background or, um, you know, other
areas. I think that the key is to have lots of different experiences. Um, I think AI allows us
to have a lot of different experiences, experiment. Um, I actually like to tell people,
you know, learn from your failures. Like I've learned the most when I have completely screwed
up. There were a few products that haven't gone so well that you haven't seen. We've learned a
ton from them. And, uh, you know, just, just keep learning. That's, that's my best advice.
So this is where I'm dangerous. Um, there is a story you told me where you was one,
one of your lessons that you learned early in your career. Do you remember what you told me?
Oh, you're going to have to, I'm dating myself. You're going to have to remind me.
Yeah. It was the one with the, with the PNL. Oh yes. Sorry. I think it's a good one. I think
it's a good one. If you don't mind briefly talking about that. Cause I think it's a great lesson
learned. I learned from it too. Well, I'll give you, I'll give you two. Can I give it to you?
You can give me two. Cause I, you know, I'll give you one. If you're an engineer and one,
if you're a business person, um, if you're an engineer, I can say for sure. The first
very, very first product I worked on, I was working at, uh, at IBM at the time,
uh, we announced the product that day and I can tell you nothing was working.
Like nothing was working. Like the chip wouldn't turn on. Um, those are scary times when that
happens. And, uh, what it allows you to do is just rally the team around. Uh, like I said,
engineers are super, um, are super innovative and creative when, where there's a will, there's a way.
Yeah. So, uh, and then the one that you're talking about, I told Ryan this story, this,
this wasn't supposed to be a public thing. I know. I told him I'm going to bring him something new.
You told me I could do whatever I want to do up here. Yeah. Okay. So the very first business I
managed, um, I was, uh, like I was an engineer turned business person and, uh, no offense,
but I thought it was going to be relatively easy. I didn't think there was going to be a big
transition between engineering and business. So for the very first business I managed,
I literally missed my first annual operating plan by 50%, like 50%. And I was devastated.
Like my boss told me, Lisa, you're terrible. You're awful. How could you do that? And I said,
you know what? He was right. I could have done better. And so, uh, I learned from that. I learned,
I learned how to ask more questions the next time. Uh, but more importantly, um, every one of these
are learning experiences. And I appreciate you sharing that because I think people don't see
that side of that. You know, we've all made mistakes. We've always, how do you learn from it
and how do you get better? And obviously you still recall it, right? It's still, yes. Remember
those to get better. Um, we've got two minutes left. So here's my thing. The future of AI is
blank. I'll give you a second to think about it. The future of AI is mind blowing, mind blowing
what we're going to be able to do with AI. You heard David Connelly say that he has been able to
create things that he hadn't expected with computing itself, but computing plus AI will
take that to an entirely new level. So, you know, and you also mentioned, you know, yeah,
you're an engineer and you've, you know, there hasn't been too many people who transitioned to
be a great leader. And I will, she's so humble. She won't let me say kind words in the back,
but I'll say kind words up here. Cause she can't do anything. So I'm still here. Um,
you wanted this conversation when we talked, when you asked me to be here,
you wanted to give something back to the audience. You wanted it to be different.
You wanted people to feel how technology innovation is, is not just the future,
but how we need to come together. So I give kudos to you and your team, allowing me
to roam around AMDs, everything that I could get my hands on the entertainment industry,
which is really huge that I think you all are changing the game, but also
for, again, I'll repeat this, Lisa and AMD, at least for me, the ecosystem that you built,
that we are together, we're building together and staying ahead of the curve is just inspiring.
Um, I think for all of us and just as entrepreneurs and folks in the room.
So I appreciate Lisa. I can't, I can't thank you enough people. Dr. Lisa Sue.
