AMD's CEO Wants to Chip Away at Nvidia's Lead | The Circuit with Emily Chang

Subtitles section Play video

There's no shortage of drama when it comes to chip making.
Even when you're winning, the game changes.
Do you ever feel like you're running a race that just never stops?
Oh, of course. Of course, absolutely.
That's exactly what we're doing.
I'm running, like, a different company every few years.
Like, the industry is changing that fast.
The technology that we're building is changing that fast.
It's a fantastic race, though, because each race is different.
It's like a different movie, but it's a more exciting movie each time.
Today is a really big day for AMD.
Lisa Su, the CEO of Advanced Microdevices, likes to be in the driver's seat.
Born in Taiwan and raised in New York with a passion for electrical engineering and three degrees from MIT,
Su worked at a couple of big-name chip makers before clinching a big victory, transforming AMD from a failing chip company into a formidable force in gaming,
PCs, and the cloud.
She's quick to remind us that her secret weapon is an obsession with the tech itself.
Her no-nonsense approach probably helps, too.
I've heard you don't take s*** from anyone.
Who's told you that?
Seriously, I've heard you're tough as nails, and I'm sure you have to be.
Well, I like to win, if that's okay.
I would tell people I was doing semiconductors, and they would say, what? Like, what's that?
Like, you build chips? Like, what does that mean?
People didn't even understand, like, where technology really fit in the grand scheme of things.
Like, now, everybody knows.
Like, everybody talks about chips.
Chips rule just about everything around us.
It's one of the most ubiquitous and consequential technologies in the world.
They power your smartphones, cars, planes, power grids, the Internet, the global economy, and even Formula One races, where AMD sponsors the Mercedes-AMG Petronas team.
I think the adaptability that you need to have in this sport is similar to a high-performance real-world company.
The technology develops, and that's why it's so much fun, and what I love is the stopwatch never lies.
Chips are at the heart of a global technology race and a battle for supremacy between China and the United States.
So for the AMD CEO, the heat is on.
She's in the middle of a race of her own, pulling off the impossible by surpassing arch-rival Intel, only to find that AI has completely changed the game, and now NVIDIA is in the lead.
I'm in Austin to see how Sue competes on and off the track and if she can pull off the unthinkable once again.
Hello. Hello.
Thank you so much for doing this.
Oh, thank you. Thank you for coming.
So the first thing we did when we got here is we went to your lab and your engineers made us this chip.
Wow.
It's our own chip.
A circuit on a circuit.
This is our epic with the circuit with Emily Chang.
You've been CEO of AMD for 10 years.
You've been working in the chip industry for decades.
But I want to go back to a moment.
When did you first hear that you were going to be CEO of AMD?
It's a phone call you get from your chairman of the board, and he said, Lisa, can you talk?
I'm like, sure.
And he goes, it's time.
And I'm like, time for what?
He said, you know, the board has made a decision.
They would like you to be the next CEO.
I said, OK.
Well, how did that feel? Because it was a tough time.
AMD was trading at $2 a share.
One guy said it was deader than dead.
What made you think you could prove them wrong?
The opportunity to be CEO of AMD, it was truly like a dream come true.
It's like, wow, OK.
Like, you think that maybe you can do this?
And here you're going to have the opportunity to lead a major semiconductor company on, yes, you're right, it wasn't the best time.
I was pretty confident that we had the right pieces.
They just needed to be put together in the right way.
AMD was founded in 1969 by executives from Fairchild Semiconductor, the famed company that pioneered integrated circuits.
Founder Jerry Sanders was famous for saying, real men have fabs.
Those are the plants that manufacture chips.
AMD started off as a memory chip maker like Intel and followed its larger rival into new markets, but always struggled to achieve the scale and performance Intel enjoyed.
With a cash crisis on its hands,
AMD had to spin off those fabs and outsource its manufacturing.
A sign of the uphill battle Sue was facing, few believed AMD would make it until she started making her moves.
Now AMD stock is trading in a totally different stratosphere.
Was there a moment along the way where you really saw AMD turn a corner?
Probably the most important moment for us was the launch of our new architecture, the launch of Zen.
You'll hear us talk about Zen a lot.
Zen was our clean sheet design to really design the next generation architecture in CPUs for the next decade.
AMD's main competition in designing the best chips comes from two players,
NVIDIA in graphics processing units, or GPUs, and Intel in central processing units, or CPUs.
A CPU is the Swiss army knife of semiconductors.
They run operating systems and programs, retrieve files, and are generally the heart of a computer.
A GPU, or graphics chip, is much more specialized.
They were designed to do multiple small calculations simultaneously.
That parallel processing makes them ideal for AI.
The new Zen architecture put AMD back on top and gave it a range of chips capable of taking market share from Intel.
You proved the naysayers wrong.
That has to feel like vindication.
Like, do you ever just, like, shut the door and go, yes, all by yourself?
I don't know that I do that, Emily.
You know, those first few years were more about just stabilizing the company and getting ourselves on a good roadmap, and then we went through a period of just tremendous growth around the portfolio, and now we're in the AI world.
This current AI wave is moving so fast.
How would you describe the pace right now?
I truly believe AI is the most transformational technology that I've seen in my career.
And, you know, we've seen all kinds of technology transitions.
We've seen the rise of the internet.
We've seen, you know, the rise of the PCs, mobile phones, the cloud.
All of these things were all really megatrends.
I think AI actually surpasses all of that.
The rate and pace of change in the industry, it's, like, faster than anything we've ever seen.
It's like we've made more progress in the last 18 months than certainly the last 10-plus years.
Now, to win and keep winning, you have to keep up the pace and keep bringing the heat over decades.
How do you keep up your own pace personally?
I really enjoy just the learning part of technology and the learning part of being in our industry.
Like, I spend a lot of time with our engineers as they're developing products.
And in every one of those conversations, I learn something.
And I learn something that I can then say, hey, you know, maybe we should do this just a little bit differently.
I come here for two reasons.
I come when there are good things.
Because I, you know, want to be able to see some of the new technology myself.
And then I come when there's a problem.
And when there's a problem, I'm here to cheer them up.
Oh, to cheer them up, not to crack the whip?
No, because at the end of the day, like, these guys love what they do.
I've heard you make some late-night trips to the lab.
I've done late nights, I've done weekends.
I don't want to surprise them.
It'd be too much of a surprise.
I can't just pop in.
You might have a heart attack.
That's right.
I'm going to show you some of our latest tech.
This is a gaming chip.
We're pretty sure when it launches, it will be the fastest gaming chip in the world.
What makes you think this is the fastest gaming chip in the world?
It has a special technology, which we call X3V, which puts, you know, memory really stacked on top of the processor.
So it's phenomenal for gaming.
The fastest gaming chip in the world.
Does it look like it?
Who would know?
She would.
We talked a lot about AI.
You need to at least see the technology.
I want to see it.
Let's take a look here.
Bring it on.
Yeah.
So what we have here is our...
These are our AI chips.
This is MI300.
Engineering sample, it says.
And this is how it looks actually on the board.
Heavy.
The weight of AI.
So this is currently in production.
One of the things that we run is, you know, sort of a lot of the AI workloads, the stuff that's running on Microsoft Azure, you know, GPT, GPT 4.0, those things they're running on MI300,
Metas, Llama.
Did I hear that you refer to these machines as your children?
They're all my kids.
Yes.
So if you were to actually step into my library at home, you would see chips galore.
Because every time I launch a chip, like I save that as a souvenir.
And there's no favorite.
Although I have to say, this one has gone a long way.
Shh, don't say it too loudly.
In this new race to deliver the best chips for AI,
Sue is focused on catching up to NVIDIA.
A niche player that got its start in the 90s making chips for computer graphics and video games,
NVIDIA realized early on that its GPUs could be pivotal for AI, and suddenly blew past its competitors when Chat GPT came on the scene.
NVIDIA's chips are the hot commodity for Meta, Google, and Amazon, sending its stock through the roof and besting Apple as the most valuable company in the world.
NVIDIA is seen as the dominant force in AI right now, and AMD is widely considered the next best option.
Is that good enough for you?
Do you like being seen as the underdog?
I wouldn't call it a like or dislike.
What I would say is, you know,
AMD has had a history of doing amazing things.
And from that standpoint, we've always been in this place where we haven't necessarily had the same amount of people that, you know, other larger companies have, but we've certainly punched well above our weight in terms of technology capability, in terms of impact on the industry.
This is the most complex, highest performance computer.
People are kind of obsessed with the fact that you and Jensen are distant cousins.
I know that you are competitors, but do you ever trade notes?
Well, first of all, Jensen is brilliant, so I fully admire him for that.
We were really distant, so we didn't grow up together.
You know, we actually met at an industry event, so it wasn't until we were well into our careers.
So no family dinners, just to put the rumor to rest.
No family dinners.
And it's an interesting coincidence.
So many of your customers are building their own chips.
Hyperscalers like Apple and Meta and Google and Microsoft and Amazon.
Is that a threat?
This is a huge market, Emily.
There's, like, more than enough chips for everybody to build.
And the beauty of this is everyone's recognized that, you know, chips are so important.
Yes, so it's natural that they're going to have their own capability.
Has the shakeup of a very established industry surprised even you?
It has a little bit, I have to say.
When you think about this conversation about the resiliency of the world depending on, you know, semiconductor diversification, like those kinds of things, those are big things.
They're beyond an industry.
They're really sort of underpinning sort of the global economy, you know, underpinning national security.
It is something that, for someone who's been in semiconductors for a long time, it's different.
This is a very complex business with trillions of moving parts.
One wrong bet or manufacturing delay can cost billions and give competitors the chance to jump ahead.
Just ask Intel.
It's just been a stunning fall from grace.
This company used to be the world's dominant chip maker.
Over the years, they have fallen away from being the leader.
Their technology is not advanced.
We saw the worst share decline in about 40 years for the company.
They gave a very grim growth forecast.
They also made or announced plans to slash 15,000 jobs.
They've always been on the brink, and they always seem to find a way, but we've never seen anything like this.
Back in San Francisco,
I wanted to get perspective on the changing dynamics of the chip industry from Bloomberg reporter Ian King, who knows the stakes better than anyone else.
Explain the pace of the chip industry.
Does it feel like a race to you?
It's absolutely a race.
It's a murderous race.
You fall behind, you are in big trouble.
For decades, Intel had more resources than anybody else, more revenue, and was way ahead.
It set the agenda.
Takes five years off, basically made some bad decisions that it shouldn't have made, and now we're in a position where they might never catch up again, right?
The train does not wait for anybody.
What do you think Lisa's legacy will be?
It's already cast in stone, right?
They're going to build a statue of her because she took a company that was an also-ran, that used to get laughed at, almost made a difference, right, AMD, and made it a significant provider of cutting-edge technology, a serious company, one that people believe in and trust, and she did that.
Are we trading the dominance of one company for another?
The direction for the industry was set by Intel for decades.
Right now, the whole AI thing is built on NVIDIA.
NVIDIA is coming out with software, new chips, new services faster than anybody else's, and everybody's just following them because that's the easy thing to do.
At a certain point, they become utterly pervasive, and then the cost of replacing them becomes difficult.
So what does AMD have to do to catch up?
Depends what you mean by catch up, right?
This industry tends to have a very strong player, number one in each market, a good number two, and then a number three where you're like, why are you bothering?
At the moment, AMD is a solid number two and doing well.
They have billions of dollars of revenue from this AI accelerator that they didn't have last year, so they're doing way better, but compared to the tens of billions of dollars that Jensen's bringing in at NVIDIA, they're a long way behind.
Let's talk a little bit about the geography.
How did the chip industry become so concentrated in Taiwan?
The industry is concentrated in South Korea, in memory, in Taiwan with logic manufacturing.
Intel used to sneer at them and say, outsourced manufacturing in semiconductors doesn't make sense, it's not practical, it'll never be as good as we are.
Guess what, it's way better than they are right now.
It's enabled NVIDIA, it's enabled AMD, and a host of other companies that could never compete with Intel to do more than compete to win because it's providing the best manufacturing in the world.
However, this has created a very geopolitically difficult situation.
Taiwan is, at its closest point, about 100 miles away from the coast of mainland China.
As you know, China regards Taiwan as a rogue province and part of its sovereign territory.
That tension has existed.
The U.S. has been the guarantee of Taiwanese safety.
With chips becoming more and more important geopolitically, and we've seen that tension arise, obviously there's concern that maybe
China takes a strategic look at Taiwan in terms of taking hold of it.
Will silicon ever move back to Silicon Valley in a meaningful way?
Will chip manufacturing come back to the U.S.?
I think we will see chip manufacturing back in the U.S.
I think we should recognize, though, that the semiconductor world is global.
Like, that ship has sailed.
Was moving manufacturing abroad a mistake?
In hindsight, I think we would have been happier if there was more manufacturing in the U.S.
American manufacturing, the backbone of our economy, began to get hollowed out.
Companies moved jobs overseas.
Today, we're down to producing only around 10% of the world's chips, despite leading the world in research and design of new chip technologies.
The U.S. government is considering capping exports of your chips and NVIDIA's to other countries.
Is that the right call?
The technology that we're building is some of the most powerful chips in the world, and so we totally get that there's a national security element of it.
We want there to be as open a market as possible while still taking into account, you know, the national security interests, and it's a give and take.
Do you worry about the future of Taiwan?
We all worry about ensuring that there is resiliency in the supply chain.
So, put China and Taiwan aside.
A few years ago, we had, you know, a big storm in Texas, and, like, things were out for, like, a few days.
You know, that's geographic concentration that you want to make sure that there's resiliency for.
You were born in Taiwan, and I know you go back to your hometown often.
What's your favorite thing to do when you're not visiting fabs?
You know, I really do love Taiwan, and I was born in Tainan, so it's a small city in the southern part of Taiwan.
Actually, I was just there not too long ago to visit family, and we have a big family.
Like, you know, my dad had, like, nine siblings, and my mom had, like, six, so it was like a big family.
So, there are lots and lots of cousins and aunts and uncles and all of that, and so, yeah, it's just fun to kind of be a part of that.
Everyone's curious about the habits of super successful leaders.
What are your go-tos?
Like, what's in your daily routine?
Well, this morning, I was up at 6.30 boxing, so that's in the daily routine.
You need to have energy when you start a morning.
Heck yeah.
And, you know, a day is either, you know, here in the office or it's on the road.
But when I'm here in Austin, it's a good day.
When the pace is relentless and perfect execution is the only option, sometimes you need to take your foot off the gas and watch another team compete for some inspiration.
Formula One is the most technologically advanced motorsport in the world, and AMD's chips give the Mercedes F1 team an edge in the competition.
Sue is giving us a front-row seat to the action because, well, she likes fast cars, too.
We have a lot of Porsches in our family.
That's our thing.
One of the things is all of my cars have, like, our product names on it, so I have a Ryzen, and I have an Epic, and I have a Radeon, and I have, like, an Epic II.
That's amazing.
Yes.
I heard there's an AMD color way.
There is a color palette.
The Epic color is graphite blue metallic, and it's a gorgeous color.
And that's the color of your Porsche.
It is.
That's epic.
When you look at a car, an F1 car, does it look like a computer on wheels to you?
It absolutely does.
It's actually even more than that.
I mean, the car itself is just one piece of the equation, right?
What's so amazing about the entire Formula 1 technology is it's the car, but then it's also just everything behind it that helps get the car ready.
I mean, it's filled with chips and sensors.
I love the fact that, you know,
Toto actually keeps us informed about how things are going, and, you know, it's great that our technology helps them make some of those adjustments.
At Formula 1, along with star driver Lewis Hamilton,
Mercedes racing boss Toto Wolff is an institution.
He works hand-in-hand with companies like AMD and a team of elite engineers to push the limits of speed and performance for their cars.
There's a million different things that go into winning.
How do you figure out the one thing that nobody else knows?
There is no such thing as one thing.
I agree with you. It's true. Absolutely.
Cheap industry or motor racing is just good engineering.
Empowerment of people, development of people, and giving them the tools to come up with good ideas and make them reality.
Speaking of the technology, how do you think AI is going to change racing?
Engineers would always say the weakness of the car sits between the steering wheel and the engine because it's irrational.
It has good and bad days.
This human interaction is difficult to replicate because it's an infinite amount of data we generate.
The driver has so many sensors.
We have hundreds of millions, if not billions, of sensors in our body that make the individual still much superior to the machine in the racing car.
So we are in our infant's shoes in terms of AI and application in Formula 1.
Lewis, how are you? It's so great to see you.
Love to see you.
What's your vibe today? What do you think?
How are you feeling?
Practice 1 was horrendous.
It was like the worst practice.
I had a big spin.
This is a circuit where all the high speed is all about confidence.
You have to get building blocks.
So when you have a spin like that, it knocks you down the ladder.
You don't have a lot of time to get back there.
So in this session, I just have to go for it.
Hope that the car is there.
And you're so involved on the engineering side of things.
What drives you to do that?
That's the fun part.
That's where you tap into your creativity.
I'm not a designer in terms of the car, but I love trying to understand it and trying to see if I can pull something out of the designers that they have already done.
By the way, Lewis is a gamer too, so he uses some of our gaming technology.
I only ever play Call of Duty.
I've been looking recently, like what else is there?
I know there's a lot that I'm missing.
The thing that there's such a nice parallel with Formula 1 is we learn in every race.
It's the same thing in our world in technology, right?
We learn with every customer environment that we put our technology in with the entire F1 season.
Like every race, you learn something and you adjust.
While it's still early days in the next great technology race,
Sue has shown she can master a pivot and deliver results, something none of her predecessors could at a company that many had written off.
And typical of Sue, she skips the victory lap and is laser-focused on the next step forward.
What else do you want to accomplish in the world, aside from making AMD succeed?
I'm passionate about the next generation of leadership, so I was very lucky in my career.
People paid attention and gave me opportunities.
I believe that's part of my job as well.
There should be more women in technology.
I think there are more.
We've made a lot of progress over the last number of years, but there's a lot more to do.
So what's the lesson in there?
Have confidence in yourself.
Take a chance. Don't worry about failure.
Sit in those meeting rooms and say what's on your mind.
And volunteer.
I was given this advice when I was a young engineer.
It was, you know, run towards problems.
And what does that mean?
It means, look, we're all going to work really hard every day.
You might as well work really hard on something that is really important.
And so look for that hardest problem to solve and volunteer to help, because you're going to learn a lot in the process.
You're going to distinguish yourself.
And what's the worst thing that's going to happen?
Like, you make a mistake. Okay, fine.
Get up. The next day is another day, and you will have learned so much in that process.
♪