Although 90% of the world's advanced semiconductor chips are manufactured in Taiwan, Taiwan's lag in quantum technology may cause it to miss out on important opportunities in the quantum era. Jackson Hu, former chairman and CEO of United Microelectronic Corp. (UMC), expressed concern over the risk, and warned that TSMC's lack of quantum technology advantage may cost its orders. He also emphasized that if Taiwan missed out the quantum era could have decades-long impact and more than hundreds of billions of dollars in lost business opportunities. The following is a DIGITIMES Asia exclusive interview with Hu (translated from Chinese):
Q: You have served in the semiconductor industry for many years, and now you are also enthusiastic about the development of quantum computer technology. Can Taiwan's success in the semiconductor industry be extended to the quantum era? China's semiconductor technology is four generations behind Taiwan's, yet it has gained an edge in quantum technology.
I was interviewed by the South China Morning Post in 2019 and they asked why China's semiconductor technology was so far behind that of Taiwan. In fact, a friend from the China Semiconductor Industry Association told me that China actually started developing semiconductor technology in the 1960s, but the Cultural Revolution persecuted many professors who were studying semiconductors. Though the Cultural Revolution ended in 1976 and Chinese college entrance exams resumed in 1977, the children who previously dropped out of school due to the Cultural Revolution had to spend 15 years to be ready for semiconductor research as they went through high school, university, and graduate school.
Richard Chang, the founder of Semiconductor Manufacturing International Corp (SMIC), is a semiconductor engineer who grew up in Taiwan. Why a Taiwanese built China's first 8-inch wafer semiconductor factory? Why not someone from the mainland? That's because China wasted a whole generation of semiconductor talent in the Cultural Revolution. But now China is second only to the US in terms of quantum patents and investment, and has gained an edge in quantum communications without the need for advanced manufacturing processes. I am worried it will have a chance to leapfrog and overtake the competition.
I am worried that Taiwan is not paying enough attention to quantum, and if Taiwan falls behind because it cannot catch up, the impact will last decades. Currently, Taiwan only enjoys a global leadership advantage in semiconductors, it is especially critical for Taiwan to be prudent because semiconductor technologies are approaching physical limits.
Q: You mentioned that you are worried about Taiwan's quantum technology lagging behind. Actually, Intel and Global Foundries have both entered into quantum hardware manufacturing, and it is said that chips based on silicon spin qubits do not need to use advanced processes at this stage. Does that mean mature-node foundry makers including UMC also have a chance to join the quantum hardware race if they commit resources to it?
Of course, the use of 28 nanometers is sufficient, but the research team has to convince the companies themselves. After all, research and development require resources and manpower. Academics need to convince the industry to become their research and development partners.
If the mature process nodes of semiconductors will become a red sea competition in the future, these companies whose technology is currently stuck in the mature process may as well start to develop a new path toward quantum research and development, but it is difficult to achieve that in the short term. However, silicon carbide (SiC) and gallium nitride (GaN) compounds are currently being used in power and microwave applications and will remain popular for a while, bringing business opportunities for mature processing nodes.
IBM is making rapid progress on the quantum front, aiming to produce a 4,158 qubit quantum computer by 2025, which is actually ready for commercialization and service. If the semiconductor industry can find a quantum hardware research team willing to collaborate, they will not need to figure it out from scratch and may be able to make some achievements in two or three years, and then they will naturally become their suppliers in the future. Therefore, I encourage these semiconductor companies to partner with quantum research teams.
The high-tech culture in Taiwan has always been more hardware oriented, and there are relatively few software developers. Nowadays, there are many teams developing quantum software all over the world, and many products have already been implemented, such as compilers, logic synthesizers, and so on. If you develop hardware, you still need a set of software to go with it.
Q: Taiwan has set up a quantum national team last year, what is the progress so far? If Intel and Global Foundries have mastered quantum technology, is there a possibility that they will take away TSMC's orders in the future?
Although the Quantum National Team was established with a claimed funding of NT$8 billion (approximately US$264 million), there is no concrete plan or action. Although I am retired and not an advisor to the national team, I am very worried to see that Taiwan is already lagging behind the rest of the world but still has no sense of urgency.
Several IT and semiconductor industry people in Taiwan said that if quantum computers are available to everyone one day, then their companies will not be absent, but they probably did not understand the significance of the quantum era. A large company with plenty of resources should set up a research team of a few dozen people and allocate funds to understand the future impact of quantum computers on their business.
Intel has developed silicon-spin technology and will launch a QPU with 13 qubits in 2023, and it also has the software capability. As soon as Intel feels it is good enough for commercialization, it may snatch the orders from TSMC.
For example, if Nvidia wants to combine CPU and QPU in the future, if I were Pat Gelsinger (Intel CEO), I would tie the two together and sell them! Or if a customer is willing to place an order for both a CPU and a QPU, I'll give him a discount. This is indeed a risk. In addition, GlobalFoundries is working with PsiQuantum because it acquired the IBM fab which happened to have developed silicon photonics technology.
Nvidia has already spotted the quantum opportunity and designed a software with a quantum interface. So it's like announcing, "I'm ready," and using the cloud to throw complex problems to do quantum computing, which is much faster than trying to solve them with its GPUs and CPUs.
TSMC does not have quantum technology yet, so they have to be very careful. Silicon photonics quantum operation does not require an absolute zero environment, but only when reading out data. Optical memory can be used with optical fiber, which makes the system much more stable and easier to operate.
Q: It seems that the blind spot of Taiwanese industry players is that they think the development path of quantum computers will be the same as that of personal computers, so they want to wait until the day when quantum computers are available to everyone before looking for an entry point. But quantum computers may not go this way, they may go to the cloud, and it's already happening.
Many applications have indeed gone to the cloud. In the past few years, I have noticed that people in the semiconductor industry always say, "It's okay, we can just do foundry work in the future because Taiwan's semiconductor industry is so powerful anyway." However, although Taiwan is four generations ahead of China's semiconductor industry, China has already made a 60-qubit quantum computer with mature process technology.
I am afraid that Moore's law will hit the wall in 3~4 years because the size of SRAM cannot be reduced significantly in the 3nm process node level just like in the past, which is a big warning sign that Moore's law will hit a bottleneck.
Q: If Taiwan misses the quantum era like it missed the Internet era, what will be the price?
If Taiwan misses the quantum computing era, it will miss a multi-billion dollar business opportunity, or even more. This business opportunity is not just about selling quantum computers or software, but also about applications. Quantum technology can be applied in a wide range of fields such as pharmaceutical R&D, materials, information security, and optimization simulation and analysis. For example, Wall Street investment bankers can use quantum computers to analyze an optimal investment portfolio for their clients and send the data to the cloud for quantum computers to compute. If Taiwan is unable to provide a quantum computing solution, it will be missing out on hundreds of billions of dollars in business opportunities.
Since it takes many years to develop a quantum computer, Taiwan would be better off finding a quantum computer to develop applications and learning the software first. There are already foreign companies using quantum computing to simulate chemical reactions, and if this progress continues, chemical talents who are not capable of applying quantum computing may fall even further behind. Taiwan has not put enough effort into the quantum field. I spoke to Science & Technology Minister Chen Liang-gee in 2018 to encourage the government to develop quantum computers, but since Chen left the Ministry of Science and Technology, there has not been a window to address the issue.
Q: The problem of talent shortage is very severe. The semiconductor industry itself is already short of talent, how should we cultivate the talent in quantum technology?
We sent many people to RCA to study semiconductor technology in the 1970s, so why not send public-funded students to schools which are well-known for their quantum research, such as MIT, University of Maryland, etc.? We need to keep up with international standards to improve our research.
If we can continue to send dozens of students to study and do research abroad, and invite them back to teach in a few years, they can become the seeds of our quantum industry.
However, currently, there are not many quantum job opportunities, so students may have this practical concern. My suggestion is to provide semiconductors and quantum technology courses together in the engineering schools so that students can take double majors and go to work in semiconductor or computer companies after graduation, or in the future when quantum computing becomes popular, they will be able to use the knowledge immediately.
Semiconductor companies have been providing scholarships to train semiconductor talents for years and offering some more scholarships to train talents in the quantum field will only be a drop in the bucket for them.
The most important thing is to start setting specific targets for quantum hardware manufacturing. I would suggest the superconductor team at Academia Sinica and the photonic quantum team at National Central University to set the goal of successfully prototyping 30-qubit quantum computers by 2030, and let the government provide funding.
Through this process, we can train many engineers to have real practical experiences. There are few engineers with basic knowledge of quantum, and even fewer have practical experience. At present, only teachers from National Taiwan University, National Cheng Kung University, Tsinghua University, and Central University have the opportunity to use the quantum machine to have this experience.
We must train a team with practical experience so that we can have the opportunity to set up a company in the future.
Bio:
Jackson Hu graduated from the Department of Electrical Engineering at National Taiwan University and earned his Ph.D. in Computer Science from the University of Illinois at Urbana-Champaign. He spent most part of his career in Silicon Valley, accumulated rich experiences in IC design. Before the introduction of EDA design tools, he designed a 16-bit microprocessor with his own hands; Hu also designed and developed a graphics gas pedal at S3, which significantly improved the graphics capabilities of personal computers; He designed and developed a GPS navigation chip at SiRF to bring the navigation function into the cell phone. After joining UMC, a major foundry in Taiwan, Hu worked with customers to make ICs smaller, faster, and more power-efficient. He is now the director and advisor of two technology start-up companies.
Jackson Hu, former UMC chairman and CEO
Photo: Jackson Hu
