How Lisa Su Turned a Dying AMD Into a $1 Trillion Giant

China Insights
283 followers
When Lisa Su took over AMD, the chipmaker was struggling with years of losses, heavy debt and a collapsing stock price. Today, AMD is worth more than $1 trillion.
So how did she pull off one of the biggest turnarounds in tech history? From the gamble on Zen and Ryzen to a controversial China technology deal and the AI boom, this is the remarkable story of Lisa Su and AMD’s comeback.
#LisaSu #AMD #Ryzen #Semiconductors #AI #Tech #Nvidia #Chips #TechHistory #Business
At this White House state dinner, besides Musk, Apple's Tim Cook, and Jensen Huang and his wife,
there were two others who sat at the main table with Trump and the Xi couple,
Lisa Su and her husband, Daniel Lin. Unlike the others, we have almost never talked about
Lisa Su on this show. What did she do to earn a seat at the main table? Why did Trump specifically
name her to be at that table? To be able to sit at that table, you have to carry as much weight
So how does this lady carry such weight? Over the last couple of days,
I've gone through her background, and after finishing it, I only have one feeling.
This person's story is far more fascinating than I originally thought.
She led a company that Wall Street had written off for dead, with a stock price that had once
dropped to $1.60 to achieve a market capitalization of $1 trillion 11 years later.
Just three days before the state dinner, AMD, led by Lisa Su, officially broke through the
$1 trillion mark, and three days later, she was seated at the main table of the state dinner.
Today, let's talk about Lisa Su.
Let's first take a look at what's been called the most awkward live TV broadcast in tech history.
This is the 2018 F1 Chinese Grand Prix in Shanghai. Before the race start,
Sky Sports commentator Morten Brundle is doing his famous gridwalk, you know,
just walking around the track with a microphone, grabbing random people to interview, and it's
being broadcast live to the whole world. At the time, AMD was one of the Ferrari team's major
sponsors, and Lisa Su, as the CEO of AMD, was there, and happened to be the first person Morten
grabbed for a chat. Who would have guessed that Morten's first question would actually be,
excuse me miss, do you speak English? Back in 2018, nobody knew who Lisa Su was.
That year, AMD's market capitalization was around $10 billion,
and today, eight years later, AMD's market cap has grown a hundredfold.
From an unknown face at Formula One to a guest of honor at state dinners,
what have Lisa Su and AMD gone through over these years?
On November 7, 1969, Lisa Su was born in Tainan. The Su family had deep roots in the area. They
were in the paper distribution business in their early years and dominated the local market for
a decade. The Xinhe Hang sign at the entrance of the old family home is still there today.
Lisa Su's father, Su Chunhua, earned his math degree from National Taiwan Normal University
and a master's from National Tsinghua University. In 1967, he attended Columbia University on a full
scholarship for his Ph.D. and later worked as a statistician for the New York City government.
In 1972, when Lisa Su was three years old, the whole family immigrated to the United States
and settled in Queens, New York. When Lisa Su was a child, she had a habit. She loved taking
things apart. Smart people like to take things apart, just like us. When she was 10, Lisa Su
did something that would make most parents' blood pressure skyrocket. She took her brother's fancy
remote-controlled car and, well, took it apart into a pile of pieces. Fortunately, she managed
to put it back together in the end. Her father, Su Chunhua, later said in an interview that he
didn't scold her for taking the toy apart, but rather bought her even more challenging toys to
play with. Her mother, on the other hand, instilled something else in Lisa Su, a business sense.
This May, when Lisa Su returned to Taiwan, she said that her earliest role model wasn't some
tech giant or business tycoon, but her mother. As for Lisa Su's mother, she was an accountant by
trade. When she first arrived in the United States, she couldn't even speak English well,
but she managed to start from scratch in a foreign land and build up a company.
What her mother often said to her wasn't, you need to study hard, but rather, one day,
you will take over my company. And this was something her mother never said to her brother,
only to her. After growing up, Lisa Su was admitted to the Bronx High School of Science
in New York City, which is one of the top public magnet schools in the United States
and has produced a slew of Nobel laureates. Later, she went to MIT. When it came time to choose a
major, she picked electrical engineering, with the reasoning that electrical engineering was
the hardest major at MIT, so she wanted to give it a try. She studied all the way from her
bachelor's to her doctorate, and by 1994, at just 24 years old, Lisa Su had already earned three
degrees from MIT. After graduation, she first went to Texas Instruments, where she stayed for over
half a year, and the following year, she jump-shipped to IBM, where she stayed for 13 years.
During those 13 years, Lisa Su completed her transformation from a scientist into a business
leader. On the technical front, she conquered the Achilles heel of semiconductors, copper
interconnect technology. What is this? We know that inside a chip, transistors need to be connected
by metal wires. Traditionally, aluminum was used, but aluminum has a problem. It has high resistance,
so when you make the chip smaller, the signal slows down. The industry knew it had to switch
to copper. Copper is such a great conductor. But copper has another flaw. During the production
process, copper atoms can migrate into the silicon, which ends up contaminating the entire
chip and causing the whole batch to be scrapped. This problem had stumped the industry for many
years, but it was during her time at IBM that Lisa Su figured out the formula, essentially,
how to make copper and silicon coexist peacefully. In 1998, IBM brought copper interconnect technology
to market, and as a result, the same design made chips 20 percent faster, instantly becoming the
new industry standard. Today, all the chips in our phones, computers, and servers use copper
connections, a technology pioneered by Lisa Su. However, even more important than the technology
were the two pieces of guidance that shaped her destiny. An IBM executive named John Kelly told
Lisa Su that she should run toward problems, run toward the hardest problems. Not walk, but run,
because that's where the biggest opportunities are, where you learn the most and where you can
set yourself apart from others. Most importantly, that's where you can grow. Most people would hear
something like that and just brush it off as inspirational fluff. But Lisa Su took it to heart,
and she has consistently put it into practice ever since. This became most evident later on
in how she saved what was the worst company in the entire industry. The second time was in 2000,
when she was helped by a mentor, a legendary figure at IBM named Nicholas Donofrio. Let's just
call him Brother Don for short. He'll have a significant role later on, so keep an eye out
for this character, folks. So this mentor, Brother Don, recommended Lisa Su to the then-CEO Lou
Gerstner, recommending her to serve as Gerstner's technical assistant for one year. Lisa Su said
that the biggest benefit she gained that year was seeing firsthand how the CEO of a major
corporation thinks. After working as an assistant for a year, Lisa Su began to take charge on her
own at IBM, starting to manage the emerging products division. However, by 2007, she made
a decision that no one expected, to leave IBM. Why? She realized that she increasingly wanted
to be the boss of a company, and IBM could no longer satisfy her. So she went to Freescale
Semiconductor as an executive, managing a billion dollar business, straightened out this originally
chaotic company, and successfully took it public. Just as she was ambitiously pushing to take
Freescale to even greater heights, one day in 2011, Lisa Su received a phone call. On the other end of
the line was Dong Ge, the mentor she respected most from her time at IBM, and the person who
had helped launch her career. At that time, Dong Ge had already retired from IBM and joined the
board of directors at AMD, where he was desperately trying to figure out how to find a lifeline for
this company that was in dire straits. He took advantage of a business trip to Austin to invite
Lisa Su out for dinner. In fact, before they even met up, both of them were feeling quite uneasy.
Dong Ge was worried that Lisa Su might still hold a grudge against him over her departure from IBM.
Lisa Su was worried that Dong Ge might blame her for leaving IBM. So, to break the ice, Dong Ge
ordered a very expensive bottle of California Cabernet Sauvignon. Once the wine was opened,
the conversation flowed, and they were pleasantly surprised to find that neither of them held any
hard feelings. After a few drinks, Dong Ge shifted the conversation, looking Lisa Su straight in the
eye with a serious expression. He knew why Lisa Su had left IBM back then. She wanted a stage where
she could truly showcase her abilities. He said that AMD had a whole host of problems, deep-seated
issues, but it also had a group of the best engineers and intellectual property that no one
else had. He leaned forward, looked Lisa Su in the eye, and said, word for word,
this place is perfect for you. You coming here would be exactly the right move. Lisa Su looked
at him, and she was tempted. This kind of stage, the most difficult and challenging one, was
incredibly appealing to her. Just like when she chose the toughest major at MIT back in the day,
this time, Lisa Su wanted to give it a shot. In January 2012, Lisa Su joined AMD. Two years later,
in October 2014, AMD announced that Lisa Su would serve as president and CEO.
She was the first female leader in AMD's 45-year history. However, the moment she took office,
she was shocked to realize that this was already a corporate corpse that Wall Street had pushed
into the morgue, reeking of despair. To understand the situation Lisa Su was in at the time,
we have to go back to the world of 2014. Back then, people weren't buying computers anymore.
When I say computers, I mean those old-school box-like machines. Sales of these old-fashioned
machines dropped from a peak of over 360 million units in 2011 to below 280 million units by 2015.
The PC industry suffered the most brutal collapse in history because everyone's attention and
wallets had shifted to another screen, smartphones and tablets. For chip companies, this meant that
the meal Intel and AMD had been living off for 20 years was shrinking. And as for this mobile
revolution, AMD missed it completely. Qualcomm, Apple with their in-house designs and MediaTek
had already carved up the market. There was simply no room for AMD. So in 2014, AMD was facing an
incredibly desperate predicament. It was clinging to a rapidly shrinking market. And within that
shrinking pool, it was still the eternal runner-up being rubbed into the dirt by Intel.
How bad was it? Between 2008 and 2011, AMD went through four CEOs. Its market share in the personal
computer chip market dropped from 23% to less than 10%. And its market share in the server chip market
fell below 1%, almost as if it had vanished from that market entirely. On top of that, after five
consecutive years of losses, its 2015 revenue was less than $4 billion, a 40% drop from its 2011
peak, and its debt had reached $2 billion. By July 2015, its stock price had plummeted to a pitiful
dollar amount. At the time, Wall Street had labeled AMD as absolutely uninvestable. Analysts weren't
discussing how much it might grow, but rather which day it would finally go bankrupt. This was what
Lisa Su inherited. Now, when an average person takes over such a mess, their first reaction is
usually to cut costs, sell off assets, preserve cash flow, and just try to survive one day at a
time, while using the time bought to focus on R&D. That's exactly what AMD's previous leaders did.
But Lisa Su didn't do that. The first thing she did upon taking office was to conduct a diagnosis,
and her findings took everyone by surprise. She believed that AMD couldn't beat Intel,
not because of poor execution, but because the architecture of AMD's CPUs was fundamentally
flawed. The previous leader's approach was always to optimize a bit more, lower prices,
and squeeze out another 5% of performance. But Lisa Su said, there's no need to squeeze anymore.
The foundation itself is crooked, so no matter how you build on top of it, it's useless.
So she decided to tear down the whole building, start from a blank slate, and redesign a CPU
from scratch. Let me tell you how crazy this decision was. Redesigning a CPU takes 3-4 years
and burns through billions of dollars in R&D costs. And this company was already losing money,
cash-strapped, with a stock price of just over a dollar, and could have gone bankrupt at any
moment. Under these circumstances, redesigning it was like a patient who was on the verge of death,
and the doctor says, I'm not going to give you a shot of adrenaline. Instead, I'm going to replace
your entire skeleton, and the surgery will take 3 years. I mean, can you imagine? Is that a decision
anyone could accept? Lisa Su didn't stop there. She lost three strategies that completely reshaped
AMD's destiny to go along with this crazy decision. First, she completely abandoned the low-end
market, meaning no more cheap products. If they were going to do it, they would focus on areas
requiring extreme computing power, gaming, data centers, and artificial intelligence,
which hardly anyone took seriously at the time. She cut all projects that were a distraction,
like mobile phone chips. The second was outsourcing the kitchen. In the chip industry,
the most money-burning part isn't the design, it's the factories. An advanced wafer fab can
easily cost tens of billions of dollars. In the past, AMD ran its own restaurant. They managed
the kitchen, hired the chefs, and bought the equipment. But now, Lisa Su said that wouldn't
work anymore. We'll only write the recipes and develop the dishes. As for the kitchen,
we'll hire the best in the world to handle it. That means all manufacturing is handed
over to TSMC. Third, she pioneered the chiplet revolution. The traditional method is to carve
a large chip out of a single piece of silicon, much like sculpting a statue from a solid block
of marble. If there's a single defect, the entire block is ruined. The larger it is,
the more likely it is to fail. Lisa Su's approach, however, is to break it down into smaller pieces,
manufacture them separately, and then assemble them like Lego bricks. This approach significantly
lowers the cost of errors and the need to scrap everything and start over.
This concept later became the industry standard. From this point on, Lisa Su bet the entire future
of AMD on this new CPU, which gave birth to the Zen architecture. If Zen succeeded,
AMD would survive. If Zen failed, AMD would go completely bankrupt. There was no third option.
During those years, Lisa Su developed a habit of training boxing with a coach every morning.
Just think about it, an Asian woman under 1.6 meters tall hitting a punching bag in the morning
and hitting Intel during the day. By the spring of 2016, Zen had been in the works for four years,
and the first physical sample chip was finally coming back from the factory, code named Zeppelin.
This chip would go on to become the Ryzen. We all know that Ryzen later became a brilliant
comeback battle for AMD, but this Zeppelin chip actually failed. Between failure and success,
there was a key decision made by Lisa Su. What kind of destiny-altering decision did Lisa Su
make? We need to introduce the industry's workflow first, otherwise the following content will be
hard to understand. Just because a chip is made doesn't mean it's ready to be sold.
The first batch of physical chips produced by the foundry after receiving the design is called the
Once the first silicon comes back, the real purgatory begins. This phase is called
Validation means taking the actual chip and testing it item by item. Can it boot up if you
plug it into various motherboards and run Windows? Will it get the blue screen of death? Will it
crash if you overclock it? You have to go through the configurations of hundreds of manufacturers
one by one. This process takes anywhere from a few months to a year. It's the longest part of the
entire project and the key factor that truly determines whether this chip can hit the market
and when it can be released. So the airship finally came back from the foundry after four years,
and the very next step is validation. Lisa Su's office is on the fourth floor, and from her
window, she can see the lab building across the way, which is where the chips are validated.
During that time, she would look over there every single day, sending texts and making calls,
just waiting for this airship to land. But the airship crashed. The chip had a defect that had
slipped past all the preliminary computer simulation tests, so it hadn't been detected.
But when they started the actual validation, they found that this chip wouldn't even boot up,
so this batch of chips was dead and unusable. What does that mean? It means this chip had to
go back to the drawing board and be remanufactured. If you're lucky, it takes a few weeks, but a few
months is the norm. Under normal procedures, all the engineers would just have to stand by at this
point, pause validation, and wait for the new batch of chips to be manufactured and returned
months later before they could start testing again. But a delay of several months meant
Ryzen wouldn't make the early 2017 launch window, and AMD's cash reserves simply couldn't sustain
them for another few months of waiting. What made it even worse was that Lisa Su was on a business
trip in India at the time, 8,000 miles away, with a 10-hour time difference. The head of testing had
sweaty palms. They never expected something like this to happen, and exclaimed that they had never
encountered such a major issue in their entire life. What should they do? When Lisa Su received
the bad news, she didn't panic, nor did she ask whose fault it was. Instead, she immediately
ordered that testing must not stop. But the problem was, the chip wouldn't even boot up.
So how could they test it? At this point, folks, imagine a film crew, hundreds of people, already
in place, ready to start shooting the next day. And then, the night before, the lead actor falls
ill. The standard procedure would be to shut everything down, have the whole crew wait,
and resume work only after the lead actor recovers. But Lisa Su's decision was essentially,
hey, we can't stop. We won't shut down production. We'll shoot all the scenes that don't require the
In terms of the chip, that meant even though it wouldn't boot, there were still hundreds of other
functions on it that could be tested individually. She called this state the Apollo 13 mode,
just like how NASA used the existing makeshift parts on the spacecraft to bring the three
astronauts back from lunar orbit. She then rushed back to the company headquarters in Austin,
and without even putting down her luggage, she went straight into the lab to rally the team,
emphasizing that failure was not an option. As a result, the team managed to bypass the
defect in that chip, which was only the size of a nickel but packed with 5 billion transistors,
and completed a massive amount of verification in just one month.
When the foundry delivered the revised chips a month later,
the verification progress was already a month ahead of schedule.
In March 2017, Ryzen launched on time and caused a global sensation.
This battle not only pulled AMD back from the brink of death, but from that day on,
no one inside AMD dared to doubt this female CEO again.
But the chip was saved. What about the money?
You have to realize, at that time, AMD's cash flow had long since dried up.
Developing the Zen architecture had burned through all their resources.
Not only had the company been losing money for 5 consecutive years,
but it was also hanging under the shadow of 2 billion dollars in debt.
Let's rewind to 2 months earlier, in February 2016, when AMD and a Chinese company called
Tianjin Haiguang established a joint venture. AMD licensed its high-performance processor
and society technology to this joint venture to develop chips for the Chinese server market.
As for the licensing fee, the Chinese company Haiguang paid AMD 293 million US dollars.
The cooperation architecture designed by both parties at the time was quite peculiar.
The intellectual property licensed by AMD excluded the most sensitive underlying core,
meaning they didn't give away their core assets. The Chinese side would take this technology,
design the top-level architecture themselves, return it to the joint venture for production,
and finally, the finished product would be sold to the Chinese company for packaging.
Going through such a convoluted process was all for the sake of the final chip produced,
which would allow the Chinese company to label it as domestically developed and independent.
So, just how life-saving was this sum of money for AMD at the time?
Before being banned in 1919, AMD had received a total of 200 million.
It wasn't exactly the kind of money that completely saved the company,
but rather, it was the life-saving cash flow AMD needed most during that year of extreme
oxygen deprivation and suffocation, while waiting for this brand new CPU to hit the market.
The kind of cash flow that convinced Wall Street that this company was still worth a lot.
For a patient lying in the ICU, this breath, this breath that keeps them alive, is incredibly
important. Three years later, in 2018, China's Higel officially mass-produced an x86 processor,
codenamed Diana. This chip is very similar to AMD's server chips.
Linux developers in the open-source community conducted tests and found that porting the kernel
code of AMD's server chips to this Higon chip required modifying less than 200 lines of code.
From the operating system's perspective, they are almost identical chips,
just with a different manufacturer code. And behind this joint venture,
it was discovered that the Chinese Academy of Sciences and Sugon were involved.
On June 24, 2019, one week before the G20 meeting between the heads of state of China
and the United States, the U.S. Department of Commerce issued an order placing Sugon,
Higon and their related joint ventures all on the U.S. export control list.
And a month before the ban took effect, Lisa Su had already drawn a line in the sand at Computex
Taipei, stating that it was a one-time, single-generation technology license with
absolutely no follow-up. Regarding this deal, critics say that at a time when China had long
been trying to break through the x86 barrier, AMD was the only company that truly pushed the
door open a crack to let China in. Thomas Hardwell's comment at the time was that
forming joint ventures with Chinese companies is nothing new, but the depth of the licensing
AMD provided was rare in the industry. The defense argues that AMD was on the brink of
death at the time. The deal was carefully structured, the most sensitive underlying
technology was not handed over, and it was a single-generation license that did not include
future updates. Furthermore, they had notified the U.S. government in advance, so AMD cannot be
Friends, how do you evaluate this deal from Bifen? If you believe that AMD was dying and
this was a desperate, life-saving move, please type agree to show your support.
If you feel that even when on the brink of death, one should not violate regulations or
license technology to a Chinese company, please type disagree to show your support.
In 2025, Haigon and Sugon announced their merger, becoming one of the most important
giants in China's server chip industry, and the technological starting point of its takeoff
originated from the technology license purchased from AMD for $293 million back then.
This part is the most complex and least discussed account regarding Lisa Su.
You could say she saved an American company from the brink of collapse,
or you could say she inadvertently handed a crucial ladder to the takeoff of China's chip
industry. The reception of the Ryzen launch exceeded everyone's expectations. AMD had
originally promised it would be 40% faster than the previous generation,
but the actual performance was even more impressive. What was even more lethal was
the price. AMD delivered the same performance for about half the price of Intel. The next few years
were a wild, compounding growth curve. In 2017, when Ryzen launched, AMD's CPU market share surged
back to nearly 11% in one go. In 2018, AMD was the best-performing stock in the global S&P 500
index. In 1919, the 7nm second-generation server chip Rome was launched, and in terms of support,
it officially surpassed Intel. By 2022, the stock price jumped from under $2 to over $85,
with a market cap topping $1 trillion. That year, AMD acquired Xilinx for $49 billion,
one of the biggest semiconductor mergers ever. That same month, AMD's market cap surpassed
Intel's for the first time. It was also this year that Lisa Su took over as chair of the board.
At the end of 2023, AMD released its AI accelerator card, officially declaring war
on NVIDIA. In 2024, Time magazine named her CEO of the year, making Lisa Su the first woman to
receive this title. In the second quarter of this year, AMD's revenue reached $11.5 billion,
a year-over-year increase of 50%, setting a new historical record. On September 21,
AMD's stock price hit an intraday high of $615, with its market capitalization officially
surpassing $1 trillion, marking an increase of over 180% this year. Three days later,
she sat at the main table of a state dinner hosted by Trump for Xi Jinping.
Twelve years ago, when Lisa Su took the helm, AMD's market cap was only about $3 billion,
and the stock price was just over $3. It fell to a low of $1.61 in July 2015.
Today, it has broken through $1 trillion. This is the greatest turnaround in the history of
the semiconductor industry, bar none. If you had bought $10,000 worth of AMD at that low point in
2015, today you would be sitting on a net profit of several million. Over the course of 11 years,
Lisa Su embodied the advice a mentor once gave her, run toward the hardest problems,
because that is the only place where you can rewrite history.
Alright, that's all for today. This piece took me a lot of time because the information was just
so vast and complex. It took me a long time to get this script organized, so I hope you all enjoy it.
If you like my show and enjoy these kinds of character stories and hall of fame features,
don't forget to give me a like. New friends, welcome to subscribe to True Viewpoint. I also
ask that you please forward and share my show with more friends and interact with me in the
comments section. Every interaction helps trigger YouTube's recommendation algorithm,
allowing my show to be seen by more people. Zhenfei thanks you all. Have a great weekend.
See you Tuesday on True Viewpoint. Bye for now.
