Source: Tonishi Business Review
Author: Tian Shanshan

Save the storm before it falls, and those who help the building will fall.

In the chip field, which is dominated by male leaders, a Chinese woman is at risk and has led one of the most outstanding corporate transformations in Silicon Valley history.

In 2014, Su Zifeng became a chip giant$Advanced Micro Devices (AMD.US)$CEO. At this point, she had only been with AMD for two years and had never been CEO of any company before. AMD is in the midst of a serious business crisis. Revenue has dropped by nearly 40%, its CPU market share has fallen by more than two-thirds, and its stock price has fallen 90% from its high point. Prior to her, AMD had replaced 4 CEOs in a row, but none of them were able to change the loss situation. Its competitor for more than 50 years, the CEO of Intel even went so far as to say, “AMD will never come back.”

Su Zifeng brought AMD back to life in just 6 years. Within 10 years, its stock price increased more than 60 times, soaring from 2 US dollars per share to about 160 US dollars today. At its peak, it reached more than 220 US dollars. The market capitalization has already surpassed rival Intel.

As a result, Su Zifeng was named “World's Highest-Paid Female CEO” by Forbes magazine for 4 consecutive years, “Entrepreneur of the Year” by “Fortune” in 2021, and “World's Best CEO” by “Barron's”.

Furthermore, AMD is a rival to Nvidia that cannot be underestimated. At the beginning of January 2019, AMD released the world's first 7nm video card ahead of Nvidia. In the CPU and GPU field, although AMD's market share is the second-largest in the industry, it is the only manufacturer in the world that makes CPUs and GPUs at the same time, and can use CPUs and GPUs on a single chip. Su Zifeng has also been dubbed the “Queen of Chips.”

Coincidentally, she and Hwang In-hoon, the founder of Nvidia, are distant relatives. Huang Renxun's mother is the aunt of Su Zifeng's mother. Therefore, some people jokingly claim that the competition between these two companies is a war between “cousin and cousin.”

Why did Su Zifeng bring AMD Nirvana back to life?

From the two sentences she said, and from her upbringing, you may be able to find some answers.

01 “Find the world's problems, stand up, and solve them”

This sentence is Su Zifeng's motto. She loved math and science since she was a child, and her parents always taught her to challenge difficult problems.

She said, “I always love solving really tough problems. This is a suggestion I got when I was young as an engineer at IBM: run for problems, because problems actually create opportunities.”

Moreover, she became interested in chip manufacturing during her college years. At the age of 17, Su Zifeng entered the Massachusetts Institute of Technology to study electrical engineering. At age 24, she received her doctorate in electrical engineering from MIT. Today, she is also a tenured professor at MIT.

“When I was an undergraduate, my first lesson was to build my own PC. I don't just need to build a circuit, I also need to program it. I really enjoyed the feeling of building my own semiconductor device with my own hands and watching how each step was completed. I love to touch and feel the products I build, which is a sense of accomplishment for me.”

“I had the chance to witness the chip manufacturing process firsthand, even though they weren't the most advanced technology at the time. At that time, we were able to build transistors on coin-sized materials and test the system, which is why I chose to enter the semiconductor industry.”

In 2014, AMD, a semiconductor company mired in a loss crisis and declining CPU market share, was not only a global problem she was pursuing, but also a once-in-a-lifetime opportunity for Su Zifeng.

02 “Walking between technology and business”

One key reason why Su Zifeng, who has never been CEO, was able to successfully solve AMD's crisis is that she not only knows technology but also has experience in corporate transformation and transformation during her career growth.

In 1995, Su Zifeng joined IBM as a technician. She helped design semiconductor chips that use copper circuits instead of traditional aluminum circuits, making the chips run 20% faster. This caught the attention of senior IBM executives. In 1999, 30-year-old Su Zifeng became a special assistant to Guo Shina, the then CEO of IBM. Special CEO support is IBM's unique method of training successors, and they can actually participate in the company's major decisions. This gave Su Zifeng an opportunity to closely observe Guo Shina's successful transformation of IBM. Through reforms, Guo Shner increased the stagnant market value of IBM nearly six times. She observed and partially participated in the entire process of transformation of large companies, thought more deeply about operating companies, and accumulated experience in saving AMD in the future.

Later, Su Zifeng became the head of IBM's R&D department and participated in the full development of IBM's Cell chip for the PS3 game console, helping IBM sign a joint agreement with Sony and Toshiba in 2001 to use IBM's chips for Sony's PS3 game console. This paved the way for Su Zifeng to lead AMD's transition to chip development in the gaming field.

It can be said that her career at IBM enabled Su Zifeng to truly experience “wandering between technology and business.”

In January 2012, Su Zifeng received an “olive branch” from AMD. Already interested in the semiconductor industry, she joined without hesitation. Afterwards, she helped AMD enter the Microsoft and Sony console markets and was promoted to positions such as senior vice president and general manager of global business. At that time, AMD's share in the PC CPU market was severely cut short. The stock price fell below 2 US dollars, was on the verge of delisting, and continued to lose money. After 4 consecutive CEOs, they were unable to turn a loss into a profit. The business that Su Zifeng is responsible for is AMD's only highlight.

In June 2014, Su Zifeng stepped up as CEO, becoming the first female CEO of AMD since it was founded 45 years ago. In the semiconductor field, most CEO positions are men. As a rare female CEO in this field, Su Zifeng has been questioned as never before.

In the beginning, she took measures similar to Guo Shner's transformation of IBM against AMD, stopping losses first. First, lay off 7% of employees, save operating costs, and prioritize “investing resources and technology in key areas that can improve profitability and drive continuous growth”; secondly, listen to customer voices to ensure good relationships with customers and partners; and most importantly, “promise to launch new products on time.”

After that, taking advantage of her knowledge of both technology and her ability to transform the company and develop business, Su Zifeng began to reshape AMD.

In 2015, Su Zifeng determined AMD's long-term focus on the high-performance computing market, and determined three directions of change around this position: building great products, deepening customer relationships, and simplifying business processes.

Instead of entering the smartphone chip market that was booming at the time, she identified three growth markets: focusing the original high-performance computing and graphics technology on gaming, data centers, and immersive platforms. These three business decisions all fit into the main positions of current consumer-grade chips — games are inseparable from expensive hardware, AI is inseparable from powerful data centers, and AR and VR are even more inseparable from immersive platforms.

She used her experience in developing game console chips at IBM to push AMD to develop chips that can be used in home game consoles, so that AMD won the top three best-selling game consoles in one fell swoop: Sony's PS4, Microsoft's Xbox One, and Nintendo's Wii U, and became a chip supplier for the top three game console manufacturers at the time. This gave AMD, which was already losing ground in the PC CPU market, a hematopoietic opportunity to continue to survive.

For a technology company, the real chance to turn around is to build a “great product.” As a result, Su Zifeng dedicated the company's limited resources to the development of Zen, a new architecture for personal computer CPUs. On March 2, 2017, AMD officially launched a CPU that shocked the industry — Ryzen (Ryzen) processors based on the new “Zen” architecture, bringing PC gamers, creators, and hardware enthusiasts around the world a CPU at the same price as Intel but with better performance. Although Intel has a 90% share of the CPU market at this time, the performance increase of each CPU generation is only about 10%.

In the third quarter of 2017, AMD received net profit of $71 million. In 2018, AMD ended a six-year loss situation and turned a loss into a profit. In 2019, AMD overtook Nvidia to launch the world's first 7nm graphics card (GPU). Also in this year, AMD's market capitalization returned to 100 billion US dollars, and its CPU market share increased to more than 30%.

In terms of data centers, Su Zifeng's main strategy is to reach new agreements with tech giants such as Google and Amazon. These tech giants need a large number of CPUs to support their exploding cloud business. In 2019, due to manufacturing delays, problems began to occur in Intel's chip processor business, and Apple decided not to use Intel chips in iPhones. Su Zifeng keenly grasped the competitors' mistakes and reached deals with cloud computing giants Google and Amazon, bringing AMD $6 billion in sales.

Su Zifeng once summed up her experience of leading AMD back to life in 5 words, “constantly fighting wars.” Now it seems that she has achieved a phased victory in the battle.

In February 2022, AMD surpassed Intel in market capitalization. This is the first time in more than 50 years that AMD has competed with Intel to surpass Intel in market capitalization.

At the same time, AMD successfully acquired Xilinx, the world's number one FPGA (programmable gate array, a typical integrated circuit in semi-custom circuits) chip manufacturer, at a high price of 49.8 billion US dollars. This is the largest acquisition in the history of the semiconductor industry. This gave AMD an additional advantage in the key data center market, and it also officially became a chip giant integrating the “second-generation CPU”, “second-generation GPU”, and “FPGA boss” labels.

Tonishi Business Review specifically sorted out Su Zifeng's 50 business thoughts in becoming an AMD leader and reshaping AMD, including 4 parts: career growth, reinventing AMD, corporate strategy, and life experience to help my readers understand the business wisdom of this “chip queen” more directly and comprehensively.

I. Career Growth

1. Great leaders are not born, they are trained, and they are also people who are willing to learn.

2. A person's success depends not only on their talents and efforts, but also on being in the right place at the right time. I am fortunate to have chosen the right path at a critical moment.

3. Encourage yourself to create opportunities.

4. Find the world's problems, stand up, and solve them.

5. I've always wandered between technology and business.

6. I never said I had to be a CEO, that's not something I wanted to think about. I've always been talking about doing what I think is important.

7. I've loved math and science since I was a kid, and my parents always told me to take on difficult problems. When I entered MIT, I had to choose between electronic engineering and computer science. Computer science only requires writing software programs, while electronic engineering requires hands-on construction. I want to build something.

8. Compared to software, I probably think hardware is more “sexy”. I had the chance to see for myself under a microscope how a chip was built on a piece of material the size of a coin. Although it wasn't the most advanced technology, it was a very precious experience for me.

9. Some people may be better at learning theory, while others are more focused on practical work. Both have their own strengths and can experience different things. I prefer to learn by doing, and I think experience is very important.

10. When I was an undergraduate, my first lesson was to build my own personal computer. I don't just need to build a circuit, I also need to program it. The feeling of building your own semiconductor device with your own hands and watching how each step is done may not be to everyone's liking, but I really enjoyed the process. I love to touch and feel the products I build, which is a sense of accomplishment for me.

11. I decided to continue my doctorate studies and focus on research in the field of semiconductors. A doctorate degree is invaluable to me. More than just an education, it was an opportunity for me to learn how to think and solve complex problems. This experience gave me great confidence.

12. I saw that the MIT PhD would be managed by the Harvard Business School MBA, and I don't think that makes any sense. I'm determined not to work for the Harvard MBA because I know technology and those people don't. (Su Zifeng's speech at the MIT graduation ceremony mentioned the career direction she chose when she first graduated.)

13. I joined Freescale (Freescale Semiconductors) as Chief Technology Officer. Over the next few years, I was ultimately responsible for running the web and multimedia business. This was definitely a choice. I wanted to drive results, and driving results required technology, but I also needed to develop the right business strategy. (In 2024, I was interviewed by the media about my career.)

14. Technology is important, but it also needs to drive commercial results, which is why I love what I do. I can work with a great team to create great technology, but also have the opportunity to drive very significant commercial results.

15. In the early years of IBM (1995-2007), my manager asked me if I wanted to be an IBM researcher or vice president. I think this is an interesting multiple-choice question. After thinking about it, I thought I might not be smart enough to be a good researcher, so I decided to work hard to become a vice president. Since then, I've had the opportunity to lead small teams that gradually grew into medium to large teams. What I love most is seeing team members come together to face challenges and complete tasks that seem difficult or even impossible. This process has always been the most enjoyable part of my career. (From 1995 to 2007, Su Zifeng worked for IBM. From a researcher to vice president of R&D, he became a top expert in the semiconductor field. This is the highest position a Chinese person can achieve at IBM. (She was responsible for designing Cell processors, which are used in fields such as PS3 game consoles and high-speed medical imaging.)

16. I walked into a room with about 25 people and I was the only woman. I'm passionate about young female engineers and want to keep them in engineering.

2. Reinvent AMD

17. Continuously fighting wars. (Su Zifeng sums up her experience at Chaowei.)

18. Many people ask me why should I join AMD? At the time, I never thought why I wouldn't join.

19. I always love solving really tough problems. This is a suggestion I got when I was young as an engineer at IBM: run for problems, because problems actually create opportunities. And I believe this is AMD's story.

20. I am a hard worker and like to get down to earth. My goal is to set lofty goals for the team and help them achieve their goals.

21. I've been dealing with semiconductors throughout my career. If you grew up as a technician, you'll know that there really aren't many large American semiconductor companies. So I'm excited to be the CEO of Chaowei Semiconductor.

22. Build great products, deepen customer partnerships, and streamline business processes. Just do these 3 things to keep things simple; if it's 5 or 10 things, then it's hard. (In 2019, Su Zifeng summed up her 5 years as CEO to reshape AMD's business priorities.)

23. For a technology company, it's important to decide what you're really good at, because you have to be the best, and you have to be the best in the industry.

24. We have large competitors, and if we can make decisions simply and quickly, we can surpass them.

25. When thinking about a product, build a blueprint for the next 5 years.

26. At the time, we would have a conversation like, “Should we enter the mobile phone field?” I answered, “We shouldn't because we're not a mobile phone company, and (and) there are companies that do a much better job than us at this. We're a high performance company, so we have to build a roadmap that leverages our strengths, which requires us to completely reform the way we design and manufacture our architecture”.

27. If AMD has any unique ideas, it is that we can put together the best processors and combine the CPU and GPU in different combinations and interconnection methods, which is actually the future development direction of accelerated computing. (In 2020, I was interviewed by the media to discuss the experience of leading AMD's Nirvana Rebirth.)

28. The world needs more semiconductors and chips. Much of what we do is standardized products for a wide range of use cases. But you'll also find some high-volume applications, such as game consoles, such as some current jobs in the cloud computing field, such as some AI jobs that will be customized. In these cases, because the capacity is very large, customization makes sense.

29. The field of gaming can perfectly demonstrate our technical capabilities, and we can also use it to enter thousands of households.

30. The future of gaming continues to be an important part of AMD's product portfolio. I do think cloud gaming is promising, but it will take years of hard work. (In 2014, Su became the CEO and chairman of AMD. One of the major strategies to revive the company was to launch the business of customizing chips for game consoles.)

31. (around 2014) Now is the time to bet the company's future on chiplet (core) technology because we can get higher performance, more cores, and lower costs. (After taking office as CEO, Su Zifeng led AMD to switch to Chiplets chip design technology and became a high-performance chip manufacturer.)

32. We have to prove that we are a great company, and I think we've done it. Proving once again my greatness and building a lasting legacy of my contributions to the world are some interesting questions for me.

III. Corporate Strategy

33. (2014) From a strategic perspective, I think we have a few choices to make. The most exciting thing at the time of 2014 was mobile phones, but we didn't make smartphones or tablets, because strategically speaking, this won't be the future the company wants to develop.

34. (2014) I judged at the time that, whether now or in the next five years, whether in cloud computing, machine learning, artificial intelligence, or the PC market, high-performance computing will be the most important demand for CPUs and GPUs.

35. AI is the most transformative technology of the past five decades, and is probably more important than the introduction of the internet. (Speech at AI PC Summit, March 2024.)

36. Whether it's the CPU, GPU, or NPU, the software ecosystem is very important.

37. The fact that Moore's Law is slowing down means we have to look for different ways to assemble chips.

38. I really believe AI will be capable of triggering growth in the PC market. Currently, AI is indeed still in a relatively early stage.

39. I think the new model running in local mode will definitely give users a better experience compared to calling (the model) on the cloud. If you use the big language model in a cloud environment, you may actually feel some delays, and if it is called locally on the PC and has sufficient computing power, it can be very fast or very private. (Speech at AI PC Summit, March 2024.)

40. The impact of artificial intelligence on our lives will continue and be profound, and all companies involved will have the opportunity to redefine the functionality of personal computers and mobile phones.

41. I'm a super supporter of open source. I don't think there is a single company that can solve all problems; companies must cooperate.

IV. Life Experience

42. Family love and education are the two most important factors that determine success in life.

43. All education is about learning to think and learn how to solve problems, whether 30 years ago, now, or 30 years from now. Technology changes over time, but if you learn how to solve problems, you have a chance to achieve something. So keep learning.

44. Learn from failure.

45. Dare to take risks and ensure that you can grow in the midst of mistakes.

46. The mentors who helped me the most were the ones who told me when I messed up. Everyone can tell you how great you are, (but) when you make mistakes, those who are really willing to help will tell you what to do better.

47. Every time you learn something new, a sense of pride develops, and this feeling will continue to accompany your growth.

48. A product roadmap is like life; the most important thing is to make choices. Choosing won't be easier just because you did better.

49. I don't have to be the smartest student in my class; as long as I focus on solving real problems, I can make a big difference. (In 2017, delivered a graduation speech at MIT.)

50. If you make your dreams a little bigger, you may actually change the world.

Reference materials:

1. “Su Zifeng: How I Reinvent AMD | “Fortune” Interview”, Fortune, 2020

2. “Chip Queen Su Zifeng and Her Half AI Jiangshan”, May 10, 2024

3. “A conversation with AMD Su Zifeng: AI is the most transformative technology in 50 years. Can it trigger a wave of PC switching?” , 2024/3

4. “How can the Chinese chip goddess break out in a man's world?” , 36 krypton, April 2021

5. “20,000 Words Complete Edition: June 6th, the latest interview with Mom Su! How to Lead AMD to Extraordinary Success, June 2024

6. “With an annual salary of over 400 million, how did she become the most profitable female CEO in the world? | [Low-profile production by Jingwei]”, November 2023

7. “Forbes cover: Chinese chip boss confronts Su Zifeng wants to take away Huang Renxun's AI crown”, June 2023

8. “Chip Wars”, Huazhong University of Science and Technology Press, Yu Sheng

Editor/Somer