From the Computer History Museum Oral History of Dr. Shang-Yi Chiang:
Intel, not TSMC, had a couple different cultures. Intel decided they want to do everything, “copy exact”. This most important principle in their R&D and their manufacturing. And what it means it's great for them. What it means, they develop this technology in R&D, use this equipment, use this recipe. They check everything very thoroughly. Everything great. Go to manufacturing, you never change it. You just follow that. Don't do any change. It's great. You have much lower risk. But the one problem is a year later, the new equipment has a better efficiency. TSMC will adapt that, Intel would not. So, then TSMC would begin to have a lower cost than Intel. So, TSMC tried to be very flexible, which it adapted.
And one reason is, obvious reason is Intel’s system. They can sell the wafer for $20,000 a wafer, because their CPU chip's very high price. And TSMC cannot sell the wafer for $20,000. We can only sell for $4,000. And we have to try to cut the cost. Next thing is Intel-- I really respect Intel. I think they are most willing to take very high risks. In every generation, they are willing to take risk to do something new. And in many key areas, for example, like high-K metal gate, strained engineering, FinFET, etc., it was always Intel, the first one to adapt that. And then TSMC would adapt in the next following generation.
So, every node, Intel's performance was better than TSMC. When I was at TSMC, I keep telling people, "We are behind Intel." We were behind Intel. "Don't look at it that you can release your 10-nanometer before Intel, then you're very happy. you are definitely wrong. Number one, TSMC's 10-nanometer definition is more like Intel's 14-nanometer dimension.
TSMC usually will wait until Intel adapted, till do it at the next following generation. Number three, not only the design rule, TSMC was also behind in transistor performance. TSMC always behind Intel’s transistor performance. And that's a good reason because Intel, their only product is CPU. And that is performance driven - they need that. TSMC doesn't need that. But usually if you try to judge who has a better technology, you go by performance. And you had to respect Intel had the better performance. When I was at TSMC, it was my last couple of years, I start an initiative. I say, "I want--" we called it Advanced Transistor Leadership, "We want to catch up and beat Intel in transistor performance." And that project failed. That project failed. I think nobody outside of TSMC’s small group knew about that. So, one of my biggest regrets in my entire career is we didn't catch up with Intel. But in my mind, TSMC technology was behind Intel. While you look at it from the surface, TSMC now is able to do 5-nanometer production. Intel's still at 10-nanometer. But I think Intel did stumble somehow. Right now, I think, Intel really is a little bit behind.
So, I firmly believe this is one of the really important reasons why TSMC succeeded. It's culture. If equipment went down, because equipment depreciation cost was so high, you really want to run your equipment 24 hours a day. In United States, if equipment went down, wait until next morning. The people come in at eight o'clock and probably go to fix it, nine o'clock. Yeah. But if at two o'clock in the morning, we just called the equipment engineer, "You come right away," he won't complain. And his wife won't complain. And that's the way it is.
Shang-Yi Chiang
Associated organizations: Taiwan Semiconductor Manufacturing CompanyFields of study: Semiconductors
Biography
Shang-Yi Chiang’s insight and expertise have transformed Taiwan Semiconductor Manufacturing Company (TSMC) from a technology follower to a driving force with one of the most advanced research and development (R&D) teams, helping it become the world’s largest dedicated independent semiconductor foundry. Known for taking well-calculated risks and making bold decisions, Dr. Chiang created an environment at TSMC for developing innovations that have made digital technology commonplace in society, profoundly impacting productivity, education, entertainment, and healthcare. Under Dr. Chiang’s direction, TSMC’s R&D organization grew from 148 people to 5,500 and has set milestones in semiconductor technology scaling at nodes from 0.25 microns all the way down to 28 nanometers.Game-changing initiatives implemented under Dr. Chiang’s leadership include a dedicated full/half node R&D roadmap, allowing customers to further reduce wafer cost. He also developed a strong lithography and electron-beam mask technology team that has advanced lithography, patterning, resist, and mask technologies for industry-leading high-density application-specific integrated circuit (ASIC)/system-on-chip (SoC) technologies for foundry customers and the logic semiconductor industry. Also important to ASIC/SoC applications has been TSMC’s high-density and energy-efficient interconnect efforts, where Dr. Chiang led his team to the industry’s first high-volume development of copper low-dielectric constant interconnects at 0.13 µm and subsequent nodes. Dr. Chiang also initiated a major direction change in three-dimensional (3D) IC technology to focus on “chip on wafer on substrate” (CoWoS) as a stepping stone to full-scale 3D-IC. This established TSMC as the leader in 3D-IC technology with industry-first high-volume production of CoWoS. This paved the way for system-level scaling for many emerging applications and has driven semiconductor industry growth.
An IEEE Life Fellow and recipient of Business Week magazine’s Star of Asia award (2001), Dr. Chiang is currently advisor to the chairman at Taiwan Semiconductor Manufacturing Company, Los Gatos, CA, USA.
Last edited:
