Three key Chinese in the AI chip battlefield: Su Zifeng, Huang Renxun and Zhang Zhongmou

Original: China Entrepreneur Magazine

Image source: Generated by Unbounded AI‌

In the battle of computing power set off by the large model, the three Chinese entrepreneurs behind it became the protagonists.

On June 14, the global AI community is waiting for a press conference. At the press conference, Su Zifeng, chairman of the board of directors and CEO of AMD, stood in the spotlight wearing a blue stand-up collar jacket and capable short hair, and intensively released a variety of new AI software and hardware products, including GPU (graphics processing Device) new product MI300X, the first initiative to challenge the king of this field Nvidia. **

The MI300X can accelerate the processing speed of generative artificial intelligence, with a memory of up to 192GB, which exceeds the 120GB memory of the Nvidia H100 chip, which means that it can train large language models with larger parameters than the Nvidia H100 chip.

While Nvidia won the vast majority of the AI computing market with its GPUs, there was always a lively debate about when AMD would launch a competitive product. Now, Su Zifeng came and set off a storm.

This storm involved the world's three major chip companies: AMD took the initiative to invite Nvidia to fight, and Nvidia's stock rose sharply that night. As AMD's manufacturer, TSMC, which is "the yellow bird behind", received a large number of orders.

The leaders of these three companies, Su Zifeng, Huang Renxun and Zhang Zhongmou, are the three most eye-catching Chinese entrepreneurs in the global AI chip industry. **This three-person game is also the epitome of the most contemporary characteristics of the AI industry.

Nvidia’s GPU is regarded as the best product for training AI large models, accounting for more than 60% of the market share; AMD is considered by US investment banks to be Nvidia’s most powerful competitor; and TSMC is the manufacturer of many chip design manufacturers such as Nvidia and AMD , its annual revenue accounts for 30% of the global semiconductor output value.

From November 2022 to the present, with the advent of ChatGPT, there has been an upsurge of entrepreneurship around the world. Behind the carnival of large models, the competition for the latest AI computing power is pushing these three companies to the center of the technology stage. **

Half a month ago, Nvidia's market capitalization exceeded US$1 trillion, becoming the world's first chip company with a market value of more than US$1 trillion; not long ago, Buffett praised TSMC at the Berkshire Hathaway annual meeting, saying that chips No company in the industry can compare with it.

The global chip industry mainly follows the following three development models: complete all links from chip design to manufacturing independently, such as Intel and Samsung; only do chip design and research and development, and manufacture is completed by foundries, such as AMD, Qualcomm, Nvidia, etc.; focus on chips Design companies complete OEM manufacturing and do not design themselves, such as TSMC and SMIC.

Su Zifeng, Huang Renxun, and Zhang Zhongmou started from different places and went to the center of the world's AI chip battlefield together.

Su Zifeng: Bringing AMD back to life

Su Zifeng has the confidence to attack Huang Renxun. In the past ten years, she has personally rescued AMD, which was once on the verge of bankruptcy, and made the stock price of Advanced AMD increase nearly 30 times in less than ten years. **

This is a company that was born before Silicon Valley. In the early 1980s, AMD began to manufacture microprocessor chips for IBM, and once surpassed Intel's own processors, accounting for about a quarter of the server chip market. However, the good times didn’t last long. When Su Zifeng took over in 2014, AMD was burdened with $2.2 billion in debt, and even had to sell its own park to lease it.

In 1969, the year AMD was founded, Su Zifeng was born in Tainan City, Taiwan, China, and immigrated to the United States with her father when she was three years old. In 1986, 17-year-old Su Zifeng was admitted to the Massachusetts Institute of Technology, majoring in electrical engineering. When someone asked her why she chose this major, she said: "Because I heard that it is the most difficult major." Su Zifeng began to study silicon technology in her sophomore year. During her university, she often went to various factories to learn and practice wafer manufacturing processes. . At MIT, Su Zifeng went all the way to Ph.D.

After graduating with a Ph.D. in 1994, Su Zifeng entered the Texas Instruments Semiconductor Manufacturing and Component Center as a technical specialist. A year later, he entered IBM's R&D department and was responsible for developing the copper wafer process. After that, he served as the head of IBM's R&D department and special assistant to the CEO, and worked in IBM for 13 years.

Su Zifeng showed her talent in the field of chip design very early on. During her work at IBM, as a researcher, she assisted in the design of a semiconductor chip that used copper circuits instead of traditional aluminum circuits, which made the chips run 20% faster, and IBM executives quickly discovered her talent. In 1999, a year after the introduction of copper circuit technology, IBM's then-CEO Louis Gerstner decided to hire her as a technical assistant.

In January 2012, Su Zifeng joined AMD and successively served as chief operating officer, senior vice president and general manager of global business. In June 2014, Su Zifeng became the CEO during AMD's reorganization, in charge of AMD's large marketing business, and became the first female CEO in AMD's history. **

Su Zifeng is affectionately called "Su Ma" by AMD engineers. On her second day as CEO, Su gave a speech to frustrated AMD employees on an all-hands conference call. "I believe we can build the best product possible," she said.

This is the first step for Su Zifeng to rectify AMD. She has already thought about the three things that must be done: create great products, deepen customer trust and simplify the company. ** "Just do those three things, just to keep it simple," she said, "because if it's five or 10, it's hard."

AMD's share in the server market continued to decline, but Su Zifeng still decided to focus most of her energy on designing chip products. She decided to prioritize the development of a new chip architecture called Zen.

This decision paid off when Zen finally launched in 2017. By the time the third-generation Zen was released in 2020, it was already the market leader in terms of speed. The Zen architecture is now the foundation of all AMD processors. Even when AMD had no chips to sell, Su spent years explaining to customers what she was doing.

Opportunities often favor those who are prepared. In 2019, Intel was in trouble due to production delays and the loss of Apple, a customer. Su Zifeng keenly seized this opportunity and led AMD to win customers such as Lenovo, Sony, Google, and Amazon, and also allowed AMD to return to the top .

In 2021, **Su Zifeng will become the first woman to win the Robert N. Noyce Award, IEEE's highest semiconductor award. **In 2022, she led AMD's $48.8 billion acquisition of chip company Xilinx, which makes programmable processors that help speed up tasks like video compression.

As part of the deal, Xilinx CEO Peng Mingbo became AMD's president and head of AI strategy. This also allowed Su Zifeng to challenge Huang Renxun with more leverage.

In June 2020, an investigation by the Associated Press showed that Su Zifeng’s annual remuneration was US$58.5 million (approximately RMB 419 million), including a basic salary of US$1 million, a performance bonus of US$1.2 million, and a stock value of US$56 million. Ten thousand U.S. dollars. ** Su Zifeng became the first woman among the most profitable CEOs in the world, and also became an indispensable key figure in the chip market. **

Jensen Huang: Entering the Trillion Dollar Club

When Su Zifeng went to the United States, 9-year-old Huang Renxun was also sent to the United States by his family from Taipei to study in a rural boarding school in Kentucky.

This was a dangerous lesson in the life of young Jensen Huang. He later recalled: "This school is actually more like a juvenile correctional school. Every child has a knife, and the students even have tattoos all over their bodies." However, this instead taught Huang Renxun to be strong and adapt to the environment.

Two years later, Huang Renxun entered a regular school to study. In addition to his excellent academic performance, he also won the third place in the national doubles in table tennis. But he loves technology more. He entered the University of Oregon at the age of 16 and studied electrical engineering. During college, he aspired to become the global graphics emperor.

After graduating from his undergraduate degree, Huang Renxun obtained a master's degree from Stanford University. Later, he entered AMD and LSI Logic as an engineer. In 1993, at the age of 30, Huang Renxun and two engineers co-founded Nvidia.

Different from chips made by Intel and AMD, Huang Renxun paid more attention to graphics chips that allowed games and images to run smoothly, because at that time personal computers had just entered the home, but entertainment functions were still missing and could not meet the needs of running games. But at that time, no one in the outside world was optimistic about the new direction of GPU.

Before 1999, Huang Renxun launched two chips. However, due to betting on the wrong technology direction, Huang Renxun exhausted the company's early investment, and the company shrank from more than 100 people to more than 30 people. At the end of 2000, Nvidia acquired its old rival 3dfx for US$110 million. In the same year, Huang Renxun challenged the chip giant Intel.

Benchmarking the "Moore's Law" of Intel founder Gordon Moore, Huang Renxun proposed "Huang's Law", that is, Nvidia's products are upgraded every 6 months, and the functions are doubled. This kind of technology update speed is faster than "Moore's Law "It's 2 times faster. **

At the end of 2006, Huang Renxun opened the GPU to software developers and launched the CUDA platform, allowing developers to use the computing power provided by NVIDIA for purposes other than graphics. While the results were modest at first, as the age of artificial intelligence dawned, developers quickly realized that GPUs are excellent at supporting the complex calculations of modern AI systems.

Later, Nvidia GPU and CUDA programming language became the basis and standard for AI development and training. This surprised even Huang Renxun. In an exclusive interview with Forbes in 2016, he mentioned that he had expected that GPUs would be used in fields other than games, but he never thought about turning to deep learning applications.

Today, the latest Nvidia products have become the hard currency in the global AI computing power. **Huang Renxun once handed over the world's first AI supercomputer DGX to OpenAI, which developed ChatGPT. Since then, the competition for the computing power of global AI companies has also become a competition for how many Nvidia GPUs there are.

In the past three years, Nvidia's market value has soared from US$150 billion to US$1 trillion. Huang Renxun is also known as the "Godfather of AI" by the outside world. According to market research firm TrendForce, among the more than 1.2 million AI servers shipped this year, 60% to 70% are equipped with Nvidia GPUs.

Zhang Zhongmou: Create 30% of the computing power in the world

In 1997, when Jensen Huang was still struggling to find a way to start a business, another entrepreneur from Taiwan, China extended a helping hand to him.

At that time, Huang Renxun had just developed the third-generation product RIVA128, which was approved by Microsoft. But the next problem before him is how to produce in large quantities quickly. At the time, it cost $100 million to build a fab, which was astronomical for a startup.

Zhang Zhongmou, the founder of TSMC, agreed to help Huang Renxun produce the OEM.

When Huang Renxun was 1 year old, Zhang Zhongmou had already obtained a Ph.D. from the Department of Electrical Engineering of Stanford University. Interestingly, after graduating, he intended to work for Ford Motor Company. Because Ford offered him a monthly salary less than that of a chip company, he chose a chip company and then worked for Texas Instruments. The next year, he was promoted to the general manager of the integrated circuit department. , and worked here for more than 20 years. In 1985, after ** failed to be the CEO of Texas Instruments, 54-year-old Zhang Zhongmou returned to Taiwan from the United States and founded TSMC. **

When he was still working at Texas Instruments, Zhang Zhongmou had a radical idea: when the demand for chips rises, chip design and manufacturing should be separated, because the companies that design chips lack the expertise to produce semiconductors. As technology advances and transistors shrink, manufacturing equipment and R&D costs will rise, and only companies producing large numbers of chips will have a cost advantage.

However, at that time, Texas Instruments, Intel, and Motorola were all developing and producing their own products. AMD founder Jerry Sanders even shouted the classic saying, "A real man needs a fab." Zhang Zhongmou's idea has not received any support. More importantly, some chip companies will worry that their design ideas and ideas will be copied.

But in the end it turns out that real men really can live without fabs.

In order to dispel the doubts of chip design companies, Zhang Zhongmou promised, **"TSMC will never design chips, it will only manufacture chips." The industry even compared this to the "Gutenberg moment" of the chip industry.

Today, none of Silicon Valley’s new wave of chip entrepreneurs owns fabs.

Chips from Nvidia, AMD, Apple and other companies are mainly manufactured by TSMC. It is precisely because of the choice of a number of chip design companies such as Nvidia that the glory of TSMC has been achieved. **In 2017, TSMC surpassed Intel in market capitalization and became the world's number one chip manufacturer. Also in the same year, Zhang Zhongmou announced that he would retire the following year and would no longer hold any position in TSMC.

"TSMC's Business Report to Shareholders" shows that in 2022, TSMC's revenue will account for 30% of the global semiconductor (excluding memory chips) output value. **In terms of shipments, in 2022, TSMC will provide 288 different process technologies and produce 12,698 different products for 532 customers.

After Zhang Zhongmou retired, TSMC took advantage of the AI Dongfeng to leave the competition far behind. His successors are increasingly aware that as the industry begins to enter the age of AI, a smarter and more interconnected world will stimulate a strong demand for computing power and low-power computing, and they will be greeted by bigger market.

References:

• Chip Wars, Chris Miller
• "Lisa Su Saved AMD, Now She Wants Nvidia's AI Crown" Forbes
• "Huang Renxun: The man who single-handedly created the "new king" of the chip market", Jiemian News