字幕表 動画を再生する
So this is Genoa, and Genoa is our latest data
center server processor.
It's actually 13 chips on one package.
At the heart of all the most advanced computers,
data centers and gaming consoles there are two
kinds of processors.
You know, we're in a world today where chips are
everywhere. They're powering everything.
And only one company in the world designs them
both at scale: Advanced Micro Devices or AMD.
It's known for computing, but now it's branching
out. Its chips are inside Teslas, the Mars land
rover, 5G cell towers and the world's fastest
supercomputer.
They used to be sort of under the covers.
People didn't realize that chips were so
important. And I think what the pandemic has done
is it's just reminded people.
And it really highlighted why chips are so enabling
to everything that we do.
AMD only has major competition from two other
companies when it comes to designing the most
advanced microprocessors: Intel in CPUs, central
processing units, and NVIDIA in GPUs, graphics
processing units.
While AMD controls far less market share than
Intel in CPUs and NVIDIA in GPUs, AMD made history
this year when it surpassed Intel's massive
market cap for the first time ever.
I think AMD is beating Intel on all the metrics
that matter. And unless Intel can fix its
manufacturing, they will continue to do that.
But a decade ago, analysts had a very different
outlook on AMD.
It was almost a joke.
For decades they had these incredible
performance problems. They just could not
execute. Every product was late or
underperforming. And that's changed.
This is the story of AMD, its remarkable comeback,
pioneering female CEO and huge bets on new types of
chips in the face of a PC slump, shifting trends and
mounting concerns around China and Taiwan, where
all AMD's advanced chips are made.
AMD is woven into the origin story of
microchips. It was founded in 1969 by eight
men, chief among them Jerry Sanders.
The famously colorful marketing exec had
recently left Fairchild Semiconductor, which
shares credit for the invention of the
integrated circuit.
Jerry Sanders was this big, larger-than-life guy.
He was a salesman at his heart, but he was one of
the best salesmen that Silicon Valley had ever
seen. Stories of lavish parties that they would
throw. And like there's one story about him and
his wife coming down the stairs of the turret at
the party in matching fur coats.
Jerry Sanders, at the time, a lot of his
philosophy was kind of new, novel.
Things like bonuses and profit sharing.
Those were not just for the executives.
AMD released its first product in 1970, went
public in 1972 and was pumping out computer chips
by the mid seventies.
It was a second source supplier for Intel by the
eighties, when Harry Levinson began his 20 plus
years with AMD.
When I first got there, most people didn't know
about semiconductor devices.
In the mid eighties, AMD and Intel parted ways.
And by the late eighties and early nineties, AMD
reverse engineered Intel's chips to make its
own products that were compatible with Intel's
groundbreaking x86 software, making PC
pricing more competitive for end consumers.
When we really got going in the x86, there was a
revolution.
AMD and Intel entered a long legal battle over the
intellectual rights to the x86 processor.
It culminated in a settlement in 1995 after
the California Supreme Court ruled in favor of
AMD's right to design x86 chips.
From there, AMD became a major player in semis,
racing Intel to become first to produce a 1-GHz
processor and making the first 1-teraflop GPU,
meaning it can handle a trillion calculations
every second.
For almost all its 53 years, AMD has been
designing chips for computers, data centers
and gaming consoles like the PS5 and latest Xbox.
But under CEO Lisa Su, it's branched out into
whole new sectors.
We're in the Tesla models S and X.
We're also in many industrial applications,
aerospace and defense applications, health care
applications.
Until just over a decade ago, AMD wasn't just
designing these chips, it was making them too.
Jerry Sanders was very famous for saying "real
men have fabs," which obviously is a comment
that is problematic on a number of levels and I
think has largely been disproven by history.
That's because as the technology advances,
making chips has gotten prohibitively expensive.
It now takes billions of dollars and several years
to build a chip fab.
And eventually it proved too much.
And when the financial crisis hit like '08, '09
AMD almost went bankrupt.
In 2006, AMD bought major fabless chip company ATI
for $5.4 billion.
Then in 2009, AMD broke off its manufacturing arm
altogether, forming GlobalFoundries.
That's when their execution really started
to take off because they no longer had to worry
about the foundry side of things.
GlobalFoundries went public in 2021 and remains
a top maker of the less advanced chips found in
simpler components like a car's anti-lock brakes or
heads-up display. But it stopped making leading
edge chips in 2018.
Unfortunately, the execution was not
sufficient at GlobalFoundries for that
relationship to be sustained.
Instead, AMD turned to Taiwan Semiconductor
Manufacturing Company, TSMC, the first chip
company to focus entirely on manufacturing.
Today, TSMC manufactures at least 90% of the
world's most advanced chips and all of AMD's.
When you think about what do you need to do to be
world class in design, it's a certain set of
skills. And then what do you need to do to be world
class in manufacturing, it's a different set of
skills. And the business model is different.
The capital model is different.
By breaking away from manufacturing, AMD
suddenly had far less capital expenditure.
Making chips has gotten so expensive because of
how precise the process is.
Now that the smallest transistors are 10,000
times thinner than a human hair.
To make leading-edge chips, 7-nanometer and
better, requires an advanced form of
lithography called EUV.
Think of it like extreme precision etching done
with a beam of extreme ultraviolet light by a
machine that costs $200 million made by ASML in
the Netherlands.
We are the only provider on the planet of this
critical technology.
TSMC was the first to deliver high volume chips
made with ASML's EUV machines, and that's kept
it at the front of the pack. But now Intel has
doubled down on manufacturing, producing
its first chips with EUV this year and committing
$20 billion for new fabs in Arizona and up to $100
billion in Ohio for what it says will be the
world's largest chip making complex.
But the projects are still years away from
coming online.
Intel is just not moving forward fast enough.
They've said they expect to continue to lose share
next year, and I think we'll see that on the
client side. And that's helped out AMD
tremendously on the data center side.
Many point to AMD's Zen line of CPUs, first
released in 2017 as the moment AMD started to
catch Intel.
This is a little hard, I'm sure, to choose among your
babies, but do you have a favorite product?
That's very hard. But if I had to choose a product or
something that was really transformational for the
company, I would say our first Zen processor.
I mean, they were literally probably six
months away from the edge and somehow they pulled
out of it. I mean, they have this Hail Mary on
this new product design.
They're still selling like later generations of
today. They call it Zen.
And it worked.
It had a massively improved performance.
It enabled them to stem the share losses and
ultimately turn them around.
Among the Zen products, AMD's epic family of CPUs
have made monumental leaps on the data center
side. Its latest, Genoa, was released earlier this
month. AMD's data center customers include Amazon
AWS, Google Cloud, Oracle, IBM and Microsoft
Azure.
And what it does is it's a very, very high
performance capability that goes into, you know,
think about it in cloud servers or in your
enterprise back end.
AMD's success at catching up to Intel's
technological advances is something many attribute
to Lisa Su, who took over as CEO in 2014.
When I started as CEO, we were probably about 8000
people and this year we're about 25,000.
So we have a few more people than we used to.
Su is the first female CEO of a major semiconductor
company. She was Fortune's number two
businessperson of the year in 2020 and the
recipient of three of the semi industry's top
honors. Su also serves on President Biden's council
of advisors on science and technology, which
pushed hard for the recent passage of the
CHIPS Act. It sets aside $52 billion for U.S.
companies to manufacture chips domestically instead
of overseas.
I was honored enough to actually be at the signing
ceremony. And when you think about sort of the
Chips and Science Act overall, I think it's a
recognition of just how important semiconductors
are to both sort of the economic prosperity as
well as national security in the United States.
With all of the world's most advanced
semiconductors currently made in Asia, the chip
shortage highlighted the problems of overseas
dependency.
And then there's the continued tension between
China and Taiwan.
I think we are all vulnerable to
semiconductor supply chain's reliance on TSMC.
If we get to the point where TSMC is cut off, if
we are cut off from TSMC, we all have bigger
problems to worry about than getting the latest
graphics cards.
So now TSMC is building a $12 billion five nanometer
chip fab outside Phoenix.
We're pleased with the expansion in Arizona.
We think that's a great thing and we'd like to see
it expand even more.
And earlier this month, the Biden administration
enacted big new bans on semiconductor exports to
China.
China has become more aggressive in what they
call their military civil fusion strategy, which is
essentially fancy talk for buying our
sophisticated chips, which are supposedly for
commercial purposes, and putting them into military
equipment to advance their military.
AMD has about 3000 employees in China, and
25% of its sales were to China last year.
When we look at the most recent regulations,
they're not significantly impacting our business.
It does affect some of our highest end chips that
are used in sort of AI applications.
And we were not selling those into China.
So overall, I would say the revenue impact has
been very small.
Something that is impacting AMD's revenue,
at least for now, is the PC slump.
During the pandemic, we pulled demand forward.
People just were at home and so we all upgraded our
computers and our gear and now nobody's doing
that again. We're all going back to work and
using work computers. W e're not buying new ones
at the same pace. And so we're just correcting for
that.
In Q3 earnings reported earlier this month, AMD
missed expectations shortly after Intel warned
of a soft fourth quarter coming up.
PC shipments were down nearly 20% in Q3, the
steepest decline in more than 20 years.
It's down, which might have been as expected, but
it's down a bit more than perhaps we expected.
There is a cycle of correction which happens
from time to time, but we're very focused on the
long term roadmap.
And it's not just PC sales that are slowing.
Many say the very core of computer chip technology
advancement is slowing down. An industry rule
called Moore's Law used to dictate that the number
of resistors on a chip would double about every
two years.
The process that we call Moore's Law still has at
least another decade to go. But there's
definitely, it's slowing down.
Everybody sort of used CPUs for everything, just
general purpose compute.
But that's all slowed down. And so now it's
something that makes sense to do more
customized solutions.
That's why in February, AMD closed on one of the
biggest acquisitions in semiconductor history.
$49 billion for Xilinx, known for its
reprogrammable adaptive chips called
Field-Programmable Gate Arrays, or FPGAs.
FPGAs are in robotics, they're in wireless base
stations that place your cell phone, they're in
cars and smart cameras or in the Mars land rover,
the Perseverance. We're in satellites in space, we're
in avionics, we're in medical systems, robotic
surgery, we're in agricultural equipment.
So very, very diverse.
Former Xilinx CEO Victor Peng has been in the
silicon industry for 40 years, including two
stints at AMD, where he now runs embedded
computing.
So this is for creating a very smart camera.
It can be changed from a hardware perspective.
It could also be changed from software because it's
got both.
An estimated 20% of AMD's revenue in 2022 will come
from FPGAs, with 72% from CPUs and 8% from GPUs.
AMD is also king in gaming processors.
While Intel controls more than 70% of the CPU market
and NVIDIA controls 87% of GPUs, it's AMD that
designs 83% of gaming console processors.
And as more and more changes come to the
semiconductor industry, Su says it's all about
being willing to reassess.
If you looked at our business five years ago,
we were probably more than 80%, 90% in the
consumer markets and very PC-centric and
gaming-centric. And as I thought about what we
wanted for the strategy of the company, we
believed that for high performance computing,
really the data center was the most strategic
piece of the business.
Earlier this year, AMD acquired data center
optimization startup Pensando for $1.9 billion.
I mean, we can quibble about some of the prices
they paid for some of these things and what the
returns will look like.
But building a custom compute business to help
their customers design their own chips, it's a
smart strategy.
That's because more and more big companies are
designing specialized chips just for their
purposes. Amazon has its own Graviton processors
for AWS.
Google designs its own AI chips for the Pixel phone
and a specific video chip for YouTube.
Even John Deere is coming out with its own chips for
autonomous tractors.
How does it feel to have some of your customers
become competitors in a sense?
Well, if you really look underneath what's
happening in the chip industry over the last
five years is everybody needs more chips and you
see them everywhere, right? Particularly the
growth of the cloud has been such a key trend over
the last five years.
And what that means is when you have very high
volume growth in chips, you do want to do more
customization.
Even basic chip architecture is at a
transition point. AMD and Intel chips are based on
the five decade old x86 architecture.
Now ARM architecture chips are growing in
popularity, with companies like NVIDIA and
Ampere making major promises about developing
ARM CPUs and Apple switching from Intel to
self-designed ARM processors.
My view is it's really not a debate between x86 and
ARM. You're going to see basically these two are
the most important architectures out there in
the market. And what we've seen is it's really
about what you do with the compute.
For now, AMD continues to advance its x86 core
computing chips while diversifying to meet the
needs of the ever-shifting and
vulnerable business of advanced semiconductors.
You really have to make big bets and kind of see
the future. What's going to happen over the next 3
to 5 years? What are the things that are going to
change? And then how do we uniquely capture those
opportunities to bring technology that nobody
else could do to the market?