The Wuxi Applied Technology Research Institute of Tsinghua University (referred to as "Tsinghua Wuxi Research Institute") was jointly established by Tsinghua University and the Wuxi Municipal People's Government in 2012. As a dispatched institution of Tsinghua University, the Tsinghua Wuxi Research Institute is a public institution engaged in scientific and technological services and managed under an enterprise-oriented model, implementing a president responsibility system under the leadership of the management committee. The institute aims to leverage the scientific, technological, and talent advantages of Tsinghua University, promote close cooperation between the university and local governments and enterprises, accelerate the industrialization of scientific and technological achievements, and foster industrial technological progress and socio-economic development in Wuxi.
Editor's Note
In 2026, as CXMT (ChangXin Memory Technologies) rushes for an IPO, China's domestic DRAM industry is ushering in a highlight moment. Tracing back to the industrial roots, the foreshadowing might be hidden in the 2008 global financial crisis. In this episode of "Chip Unveiled · Experts Talk Chips," we have an interview with Dr. Zhou Dejin, Director of the IC Innovation Service Platform at the Tsinghua Wuxi Research Institute, who is a firsthand witness to this industrial transformation. With over 20 years of industry experience, Dr. Zhou started from a frontline technical testing position, coordinated the national major science and technology special project "Core Electronic Devices, High-end General Chips, and Basic Software Products" (Core High Base) at the age of 27, and eventually transitioned into a platform operator for a local research institute of Tsinghua University. He will deconstruct from a frontline perspective: Where did the industrial talents go after Qimonda's bankruptcy? Is the cooling of semiconductor investment a good or bad thing? And, can AI truly help China's chips "overtake on the curve"?
Welcome to the Podcast Dialogue
*The following is the text transcript of the podcast, confirmed and authorized by the guest.
The Global Economy Collapsed, but China's Memory Sector Rose
Huan Shi: Dr. Zhou, you hold multiple identities as a Ph.D. from Fudan University, an industry professor, and the secretary-general of an academic society. In the public's inherent perception, scientific researchers should focus on academic research, yet you have been active on the industrial frontline for a long time. I would like to ask, is this career path balancing both academia and industry the result of your active planning, or driven by the development trend of the industry?
Zhou Dejin: Indeed, many people feel conflicted when introducing me, saying I have too many identities and don't know which one to introduce. I always tell them to just choose based on the context of the introduction, without being rigid.
My scholar identity, from the perspective of studying, was actually "forced." After graduating from university in 2003, I joined a company, TOSHIBA, mainly doing testing. At that time, I felt this position had limited technical depth, but if I wanted to transfer to chip design, the industry generally required at least a master's degree. So I worked and studied at the same time, and finally got admitted to the 58th Research Institute of CETC (China Electronics Technology Group Corporation) for my master's degree, majoring in the DSP direction.
DSP (Digital Signal Processor): A dedicated microprocessor with a special hardware architecture, specifically used for fast and real-time implementation of various digital signal processing algorithms. Its core features include independent program/data buses, hardware multiplier units, and pipelined computing mechanisms, which significantly improve the efficiency of convolution, matrix, and spectrum calculations, making it the core hardware carrier for algorithm implementation.
DSP is actually the technical predecessor of today's AI chips. Early CPUs only had addition arithmetic units, while DSP added multiplication arithmetic modules on this basis; in subsequent iterations, it further integrated convolution and matrix arithmetic units, eventually evolving into today's AI chips. It can be said that 20 years ago, I was one of the earliest practitioners in China involved in the R&D of algorithmic chips.
2008 DSP Chip Standard Implementation Circuit
Huan Shi: When you started your master's degree, you chose such a difficult technical direction as DSP? But this is also the sector most closely related to "intelligence," which aligns with your original intention of pursuing technical depth.
Zhou Dejin: Yes. After graduating with my master's degree in 2008, I joined the German memory enterprise Qimonda. This professional experience is quite story-like. I worked at Qimonda for about a year when I encountered the 2008 global economic crisis. This European memory giant ultimately collapsed and declared bankruptcy.
At that time, my workplace was the Suzhou R&D Center. Our team gathered technical experts from the US, Germany, and the Xi'an R&D Center, along with a group of fresh graduates. Later, Qimonda selected some R&D personnel to go to Xi'an to complete the project, and I was one of them. I continued working in Xi'an for another 4 months.
Huan Shi: Qimonda was highly renowned in the industry back then, gathering a large number of top technical talents. I imagine that during team collaboration, everyone could gain substantial professional accumulation.
Zhou Dejin: Exactly. At that time, there were five leading global IC memory manufacturers: Samsung, SK Hynix, Elpida, Qimonda, and Micron. Qimonda's technology was very advanced; the only drawback was the high manufacturing cost brought by its process route. But Qimonda had a distinct trait—it was candid and open to people, which might be the cultural background of German enterprises.
Zhou Dejin: During my time at Qimonda, various process solutions and technologies related to power consumption optimization were shared completely, because we were all one team with a common goal. Even after so many years, after Qimonda went bankrupt and everyone went their separate ways, I still keep in touch with many former colleagues. Everyone agrees that the team atmosphere at that time was excellent, and it was the most harmonious company I collaborated with in my career.
Huan Shi: So failure also has its value; the time for industrial trial and error was not wasted. The intellectual property left by Qimonda is still shining in the industry today.
Zhou Dejin: Its core technical heritage is now landing at CXMT (ChangXin Memory Technologies) in Hefei. Some time ago, Huang Xiaodong, Vice Dean of the School of Integrated Circuits at Southeast University, who was also an old colleague from Qimonda, shared his experience of visiting CXMT with me: when the employees at CXMT heard he was from Qimonda, they still looked up to him with respect, after all, he is considered an industry senior.
Objectively speaking, the biggest opportunity brought to the domestic memory industry by the 2008 global economic crisis was providing a foundation for the start of China's memory industry, which also paved the way for the subsequent R&D of HBM.
Huan Shi: Few people associate the development of domestic memory with the economic crisis back then. The argument you raised is very unique; thinking about it carefully, it makes sense. If it weren't for this industry shift, how would the domestic memory track have had the opportunity to enter?
Zhou Dejin: Right. Of course, this opportunity is also inseparable from the forward-looking layout of Mr. Zhu Yiming. Back then, not all enterprises saw the potential and seized the patent assets of Qimonda. Vision and decisiveness were also crucial.
After Changing Positions, Managing the National "Core High Base" Project at 27
Huan Shi: Where did you go after leaving Qimonda?
Zhou Dejin: After Qimonda went bankrupt, its related assets were acquired by Shandong Inspur, and Shandong Huaxin was established. Many former colleagues joined Huaxin; other talents flowed to Huawei and ZTE. When Qimonda announced its dissolution, ZTE directly set up tables at the factory gate to recruit, almost openly absorbing all technical personnel. There was basically no interview; you could just handle the onboarding procedures on the spot.
At that time, because my family was in Wuxi, I planned to return to Wuxi for development. Just then, there was an opportunity: my supervisor from when I was studying for my master's at the 58th Research Institute started a business, taking the lead in founding CKS (China Key System), and served as the first general manager of CKS. So by a twist of fate, I returned to Wuxi in 2009 and joined my supervisor in the startup work of CKS. At the beginning of the company's establishment, there was only 100 million RMB in cash and no other business reserves. I was employee No. 008 of the company.
In the early stage of joining, I told my boss directly that I wanted to do projects. The boss said we just started and have no projects to do; you can help connect government resources, purchase materials, and write various declaration materials for the group. So I started doing these things. This experience became a major turning point in my career: from then on, I no longer did technical R&D work, but turned to industrial management affairs.
As a secondary sub-group under CETC, CKS undertook the group's major layout plan for the IC industry. At that time, the chairman of the company was Wang Zhigang, who later became the Minister of Science and Technology. My work was divided into two major parts: first, coordinating the overall restructuring plan of the group's IC business; second, managing projects. Right from its establishment, CKS undertook the national key high-end science and technology project "Core Electronic Devices, High-end General Chips, and Basic Software Products" (Core High Base) special project. While I was responsible for the related project management work, I also had to connect with local governments to carry out various collaborative work.
National Medium- and Long-Term Program for Science and Technology Development (2006-2020)
Huan Shi: The Core High Base special fund is huge in scale, and it is the most cutting-edge science and technology project in China. How old were you when you were responsible for this work?
Zhou Dejin: Assisting in managing the Core High Base project was in 2009. I was 27 years old, and thinking back now, it still feels incredible. Placed in today's context, it is hard to let a young person in their twenties coordinate a national-level scientific research project worth tens of millions of yuan.
But at that time, talents with both an IC professional background and familiarity with the special project declaration process were extremely scarce. Coupled with the shortage of manpower in the initial stage of the company, I had the opportunity to take on the heavy responsibility. During my tenure at CKS, I comprehensively came into contact with project management, enterprise operation, and even industrial external investment planning. I had almost no experience when I first took over various tasks, but I felt I had strong learning ability and execution. In the early days, to rush the preparation of project declaration materials, I directly rented an apartment next to the company and lived there. I arrived at the office after 6 a.m. every day and left at 11 p.m., with a daily working time of up to 16 hours. Only through long-term high-intensity investment did I gradually rationalize various businesses.
In the course of work, I also clearly realized that my systematic knowledge reserve in enterprise operation and industrial management had shortcomings. Since industrial management requires continuous supplementation of business cognition, I decided to further my studies systematically and applied for an MBA.
Zhou Dejin: Studying for an MBA brought me very significant improvements. I didn't just complete my studies purely; instead, I studied targetedly with the confusions I encountered in my work. After graduation, I had a completely new understanding of the entire industrial management and enterprise operation, and could clearly judge the core goals of each task I did, as well as its positioning in the entire industrial system.
I still remember when I switched from a technical post to administrative management, a senior schoolmate told me something that left a deep impression on me to this day. He said that members of a project team in the institute were still working overtime every day on R&D, not knowing that the project would be terminated before the end of the year, and all their efforts might be in vain. Whereas doing administrative management work can clearly distinguish project priorities and avoid meaningless R&D investments.
Huan Shi: In other words, you can screen and retain the most valuable R&D projects.
Zhou Dejin: Right. After leaving CKS, I joined Centec Networks, which is now listed on the STAR Market. It is the only listed company in China focusing on Ethernet switching chips. I worked at Centec for 4 years, and then moved to Wuhu, Beijing, and other places for development, until I returned to Wuxi in 2019 and joined the Wuxi Applied Technology Research Institute of Tsinghua University. In 2019, the research institute was preparing to build an IC innovation service platform. In the early stage, there was no supporting equipment, but we still positioned it as a laboratory with high standards. Since we were going to build a public scientific research carrier facing the industry, professional depth and academic accumulation were crucial, so I started pursuing my Ph.D. in the same year.
Huan Shi: You have never interrupted your learning.
Zhou Dejin: Yes, lifelong learning is my motto.
Resolving the Pain Points of Industry-University-Research Transformation: Starting from Turning Off the Filter
Huan Shi: Your current identity is the Director of the IC Innovation Service Platform at the Wuxi Applied Technology Research Institute of Tsinghua University. How do you position the core niche and value of this platform in the entire industrial chain?
Zhou Dejin: The core positioning of this platform is a one-stop industrial innovation service platform facing the fields of IC, new energy, and new materials. Our core value is to focus on the enterprise R&D stage and provide all-round support for technological innovation and entrepreneurial landing for various material innovation subjects. I often mention a sentence in my work reports, which is also the original intention of the platform: to walk alongside entrepreneurs and innovators who possess a great vision.
Huan Shi: So it is not a single industry-university-research achievement transformation platform in the traditional sense?
Zhou Dejin: To be exact, that's right. The Tsinghua Research Institute itself has a mature achievement transformation path. Relying on Tsinghua's faculty and alumni resources, it undertakes a large number of functions for teacher-student entrepreneurship and the landing of scientific research achievements. Our innovation platform's core is to provide all-round technical support for this achievement transformation system.
Conventional sci-tech innovation carriers mostly focus on basic supporting services such as site landing and financial matching. Our differentiated advantage is to delve into the industrial frontline, providing precise empowerment for entrepreneurs and innovative enterprises from the core links of technological R&D and industrial landing.
Huan Shi: Tsinghua has top-notch scientific research accumulation and a massive reserve of intellectual property. But even for top universities, there are still many scientific research achievements and patent technologies shelved and difficult to land. In your long-term process of connecting with universities and serving the industry, how do you think we can break the transformation barrier between universities and the industry, so that scientific research achievements can truly serve the industry?
Zhou Dejin: This is a pain point that universally exists in the industry. The core reason why many university patent technologies are shelved is that there is a huge gap in value cognition between universities and the industry/enterprises.
First, university teachers and scientific research teams generally overestimate their independently developed patents, technological innovations, and project achievements.
Huan Shi: Their own achievements are like their own children.
Zhou Dejin: Exactly. Their own children are excellent, the best-looking, and they will instinctively beautify the child's real performance. But the evaluation logic of enterprises is completely different. In the eyes of entrepreneurs, from a laboratory technology to finally iterating into a mature product, achieving scaled output value, and completing commercial landing, it is an extremely long chain. The technological breakthroughs and process innovations developed by universities are often just the beginning of industrialization, which may only account for 5% or 10% of the entire commercialization work, just the first step of the Long March.
University scientific researchers regard the technological breakthrough at the starting stage as a complete entrepreneurial achievement, while enterprises see the subsequent long, high-cost, and high-risk engineering landing process. This is the most core cognitive divergence between the two parties.
In addition, the core demands of both sides are completely misaligned. University scientific researchers value academic values such as paper publication, patent authorization, and professional title promotion more; while enterprises value product landing, capacity increase, output value growth, and commercial profitability more. The two sets of evaluation systems and value orientations are completely different.
Huan Shi: One side pursues technical height and academic breakthroughs, while the other pursues industrial scale and commercial value. The difference is still very big.
Zhou Dejin: Exactly. At the same time, I also summarized a rule: short-cycle, formula-based, and lightweight technological achievements are the easiest to transform in industry-university-research; while long-cycle, heavy-engineering, and heavy-iteration product and equipment technologies are extremely difficult to transform.
Let me give a typical case. In Yixing, there is a semiconductor enterprise mainly engaged in CMP semiconductor slurry and grinding consumables. The core of this technology lies in the mature formula, which is "Know How." Just like Coca-Cola, once the formula is finalized, the subsequent production landing and product delivery processes will be very simple, enabling rapid mass production and commercialization. The scientific research achievement used by this enterprise is not from Tsinghua, but from a teacher at Hebei University of Technology. After landing, it quickly cut into the supply chains of head enterprises like Zhonghuan, and the output value grew rapidly with extremely high transformation efficiency.
CMP: Chemical Mechanical Planarization. The core wafer planarization process in IC manufacturing, relying simultaneously on chemical corrosion and mechanical grinding to globally polish the uneven thin films and metal wiring layers on the silicon wafer surface, eliminating the height difference on the wafer surface and achieving global flatness. It is an indispensable key process in advanced process multi-layer wiring technology.
Zhou Dejin: Looking at single process technologies like GaN (Gallium Nitride) and SiC (Silicon Carbide), the industrialization path is clear and the landing cycle is controllable. Many such projects in China have successfully achieved university industrialization, giving birth to a batch of high-quality enterprises. For example, Shandong Tianyue and Shandong University, Tianke Heda and the Institute of Physics of the Chinese Academy of Sciences, are all benchmarks of deep cooperation between universities and enterprises, smoothing the industry-university-research landing.
But chip products and high-end semiconductor equipment projects are completely different. Such achievements often require long iterations, debugging, verification, and engineering optimization, with an extremely long landing cycle. This creates a core contradiction: university scientific researchers need short-term output of academic achievements and completion of professional title assessments, while enterprises need to see product landing and commercial returns in the short term. The value realization cycles of both sides cannot match. Unless this cycle is stretched very long, it is unrealistic to realize the value demands of everyone in a short time. This is also the core reason why many enterprises are optimistic about university patent technologies, but ultimately fail to reach cooperation after truly communicating with the teachers.
Huan Shi: Then the innovation platform we are building now is precisely to solve this kind of industry-university-research transformation pain point?
Zhou Dejin: Exactly, it should be said that this is the core positioning of our platform. The reason it is called a platform is because there is an essential difference between our platform and third-party testing institutions.
First, our platform can provide one-stop full-chain services. Although our platform is equipped with tens of millions of yuan of professional equipment, we do not use the equipment to do R&D work ourselves. Instead, for the personalized demands of a single enterprise, we provide customized assistance; for the common public pain points of the industry, we will precipitate and standardize them into normalized services, empowering all sci-tech innovation subjects in the industry. Of course, product sales and such businesses are not within our service scope, but for the vast majority of difficulties encountered by enterprises in the process of R&D, landing, and iteration, we will do our best to assist in solving them.
Second, our platform can provide financial support to lower the entrepreneurship threshold for university teachers and students. Currently, our platform is jointly built and operated by the local government, the Tsinghua Research Institute, and our own team, possessing public attributes. Our core mission is not profit, but to land the industrial value of the Tsinghua Research Institute and the government, enabling enterprises to obtain technical support and industrial resources at lower costs and higher efficiency.
When many Tsinghua-affiliated scientific research teams and university supervisors settle in our platform to start businesses, they do not need to build a team from scratch and equip R&D resources. The complicated preliminary R&D supporting, resource integration, and landing docking work can all be undertaken by the platform, allowing entrepreneurs to focus on core technological R&D and innovative breakthroughs.
Huan Shi: It's equivalent to making up for the shortcomings of startup teams starting from scratch, solving the core problems of lacking people, resources, and supporting facilities in the early stage.
Zhou Dejin: Yes. Like a One-Person Company (OPC), in the past, he might have to lead a complete team to tackle the whole process; now relying on our platform, it's like being equipped with an AI supporting system. He can come by himself and complete the vast majority of landing supporting work, greatly reducing the cost of entrepreneurial trial and error.
Huan Shi: Your kind of deep industrial service requires extremely strong industry accumulation. It requires not only accurately interpreting the core needs of enterprises and scientific research teams but also accurately matching upstream and downstream resources and connecting with high-quality supply chains. The threshold is actually very high.
Zhou Dejin: Yes, there are just two core points: understanding their needs and efficiently integrating industry resources. This is also the core competitiveness of our platform.
Huan Shi: So your platform is essentially more inclined to be a venture capital incubation service institution.
Zhou Dejin: It can be understood this way. As a platform, we also provide a lot of support for enterprise investment. And because we are rooted in the industrial frontline all year round, accompanying startup enterprises to grow for a long time, we have the most real and thorough understanding of the enterprise's technical strength, team capability, and development potential. Many investors come to me for advice on corresponding enterprise investment suggestions.
If an enterprise continuously needs the platform's technical supporting and resource services, it indicates that its R&D iteration and business scale are steadily growing; conversely, if an enterprise detaches from the platform's services, it basically means its development has entered a bottleneck period, or is even gradually declining. We can intuitively perceive the development pulse of industrial enterprises.
Huan Shi: You are like the ducks that first sense the warming of the spring river water, able to capture industry trends at the very first moment.
28nm Has Stood Firm, but We Must Accelerate in the AI Era
Huan Shi: In the past two years, the industry generally believes that chip investment is not as hot as before, and the industry bubble has burst. Do you agree with this view?
Zhou Dejin: I think the current semiconductor industry has completely bid farewell to wild growth and entered a new stage of rational, steady, and structured development. I have experienced many troughs and booms in the semiconductor industry. When I just graduated from university, the income of R&D personnel was generally low. 15 years ago, the STAR Market had not yet been opened, and the threshold for enterprises to go public was extremely high and the cycle was extremely long. Industry practitioners mostly relied only on salary income, with almost no opportunity for equity appreciation, and the overall development was very silent.
When the STAR Market landed, coupled with overseas technological blockades, the demand for independent and controllable domestic industrial chains exploded. Coupled with the drive of leading enterprises like Huawei, the semiconductor industry quickly recovered, and industry salaries and market heat increased significantly. But what can truly bring wealth growth is probably still equity investment. During the pandemic, global supply chains were broken, and the industry faced a comprehensive chip shortage. Industrial heat reached its peak, and the market once fell into madness. At that time, as long as many enterprises were labeled with microelectronics, IC, or semiconductors, even if they were just a startup team of two or three people, they could easily get financing. It was just like the crazy atmosphere of everyone speculating in stocks in 2007; even the aunties in the vegetable market knew to buy stocks, and everyone thought they were stock gods.
But this irrational heat didn't last long. Starting in 2022, the macroeconomy began to decline, and the semiconductor industry entered an inventory digestion cycle. The three-year-long inventory digestion process caused the entire semiconductor industry to cool down rapidly, leaving a large number of investment institutions with no high-quality projects to invest in.
Huan Shi: I know, at that time, some investment institutions didn't even let their employees travel on business because there was no travel expense reimbursement.
Zhou Dejin: Because there were not many enterprises to look at either. But this round of reshuffle is not a bad thing; in essence, it is sifting the sand to get the gold and eliminating the false to retain the true. A batch of speculative enterprises that relied on the wind to arbitrage and had no core technology successively dropped out, closed down, and exited; while practical enterprises that delve into technology and focus on products have remained stable and continued to develop despite the cycle fluctuations.
Currently, the industry has walked out of the trough. There is neither the crazy overheating of the past few years nor a deep winter. The overall situation is in a stable and rational development stage. The market no longer looks at the semiconductor industry in general, but looks at it structurally. The heat of frontier tracks such as AI chips and advanced processes continues to rise, while tracks like power semiconductors and traditional analog chips tend to be stable. Capital and resources are more concentrated in high-growth, high-potential quality tracks. This is a very healthy industry ecology.
Huan Shi: This is also the result of natural market screening. Returning to the industrial chain itself, from your perspective of delving into the frontline, what stage has the independent and controllable level of the current domestic semiconductor industrial chain reached?
Zhou Dejin: Overall, it can be summarized as partially independent and controllable, with shortcomings still existing in key fields. Many people's perception of industrial chain blockade started with the sanctions on Huawei, but I personally felt the pressure of technological restrictions as early as 2014.
At that time, the Snowden incident broke out, and domestic communication chip enterprises like ZTE and Centec had already encountered overseas technological restrictions, just that few people in the industry knew about it. Because I was at Centec then, and Centec and SMIC (Semiconductor Manufacturing International Corporation) had a little business intersection, we were very clear about the entire internal sanction situation (it wasn't called an embargo at that time). Over the years, overseas embargoes on us have been repeated, and the core purpose is to contain the comprehensive breakthrough of domestic semiconductor processes.
But it is precisely the external pressure that has forced the domestic industry to accelerate its independent breakthrough. Taking Wuxi as an example, because the core of process technology is highly related to equipment, Wuxi's semiconductor equipment enterprises have risen rapidly and their technology has become increasingly mature. Companies like EWON and Shangji are doing very well. If it weren't for the embargo, their development might have been restricted to a certain extent, because most of their equipment might still come from abroad. Moreover, domestic leading enterprises like NAURA have continued to expand in scale. I think that in the field of mature processes, domestic semiconductor manufacturing processes have achieved a certain degree of independent and controllable capability.
Of course, in the fields of advanced processes such as EUV (Extreme Ultraviolet) Lithography machines and high-end photoresists, we still have a long way to go in tackling key problems. Especially, lithography machines and photoresists are complementary and deeply bound, which is a typical "chicken or egg" closed-loop problem. It is difficult for a single link to break through independently. But we can only move forward step by step.
Although I don't quite like many people on the Internet exaggerating our breakthroughs in the semiconductor field, we should not underestimate ourselves either. Objectively speaking, countless domestic engineers are rooted in the frontline and silently tackling key problems, and the industrial progress is obvious to all. So fairly speaking, in the field of 28nm mature processes, we already have basic independent and controllable capabilities.
Huan Shi: You are still relatively cautious and objective. This also means that our industry is far from the stage where we can lie flat. There are still a large number of core problems to be solved.
Zhou Dejin: Not only can we not lie flat, we even need to switch from "walking steadily" to "accelerating the run." Because the current industrial competition pattern has ushered in brand-new changes, and AI is leading a new wave of industrial development speed.
The semiconductor industry development process of the past 30 years may be quickly completed in the next 5 to 10 years. The explosive iteration of the AI industry in the past three years has a speed far exceeding the past rhythm of the industry. It feels like we have experienced 10 years of development, but it's actually just three or four years. At the same time, AI chips have put forward extremely high requirements for advanced process technologies such as 1.6nm and 3nm. In the future, AI chips are expected to occupy half of the global semiconductor industry.
Different from the previous pattern where special semiconductors such as power devices, RF, and analog chips blossomed everywhere, advanced processes and high-end chips will become the core track of industrial competition. This also means that our pressure to catch up has suddenly increased. For example, we need to catch up with South Korea's HBM technology, Taiwan, China's advanced manufacturing processes, and also catch up with the complete AI chip ecosystem built by enterprises like NVIDIA. Using the previous development rhythm can no longer keep up with the global iteration speed; we must comprehensively accelerate and strive to catch up.
Huan Shi: Every generation has its own challenges. The changes in the AI era indeed make us feel overwhelmed. In the past, we always felt that we could catch our breath once we conquered the 28nm process.
Investing Own Money in Enterprises, Betting on Domestic Chips Overtaking in 20 Years
Huan Shi: Dr. Zhou, you have been dealing with chips for over 20 years. Whether it was doing R&D in the early days, providing services later, or building a platform now, you have always been deeply engaged in the semiconductor track. What is the core vision that supports your long-term deep cultivation?
Zhou Dejin: I have a good personality and like to teach others; taking pleasure in helping others is my fundamental joy. This is also the reason why I took the lead in creating the innovation platform and the Wuxi Integrated Circuit Society. My ultimate vision is to rely on platform resources and industry resources to continuously empower IC industry technological innovation and the growth of entrepreneurs.
In the next 3 to 5 years, I hope that under our all-round support, we can cultivate three to five high-quality sci-tech innovation enterprises to successfully land in the capital market. Of course, IPO is not the end of an enterprise, but a brand-new starting point for the scaled development of the enterprise. But this can also prove that we have made outstanding contributions in this process. This is the industrial service value of our platform.
Huan Shi: Can all startup enterprises in Jiangsu come to you? Especially when there is good technology urgently needing a place to land, coming to you can not only get technical support but also connect with industrial resources, avoid entrepreneurial risks, and sort out development paths.
Zhou Dejin: Yes. We warmly welcome all high-quality sci-tech innovation teams to come and land here for development.
Currently, my daily core work is to continuously connect with and serve various sci-tech innovation enterprises, assisting teams in solving various complicated problems in the process of R&D, landing, and development. I will also personally invest in high-quality startup projects and go all out to accompany the growth of enterprises.
Huan Shi: It is very admirable that Dr. Zhou invests his own money in enterprises. Everyone is adding bricks and tiles to the domestic semiconductor industry in their own ways.
Zhou Dejin: We hope to see China rank among the top in the global semiconductor industry within our lifetime. I firmly believe that this day will come within 10 or 20 years.
Huan Shi: It is very much worth looking forward to. If we can really become the most top-notch country, most of the chips in the market at that time might be made in China or contain elements of Chinese IP.
Zhou Dejin: Yes. The rapid development of AI also provides us with a brand-new opportunity to overtake on the curve. A few years ago, EDA tools were comprehensively blocked by overseas countries, but with the deep integration of AI and EDA, a large number of high-quality EDA startup enterprises have emerged in China. After all, EDA is essentially AI, and the application ability of us Chinese is extremely strong. When a technology is no longer basic technology R&D but becomes an application development, Chinese people will walk very fast.
For example, in the past, lithography was just lithography. Now, when AI is integrated into it to become computational lithography, it starts to lean towards application.
Huan Shi: We are back to our familiar track again. Maybe we can solve the lithography application problem in a different way.
Zhou Dejin: Yes, just like NEVs (New Energy Vehicles) achieving overtaking by changing tracks.
Huan Shi: Then let's exercise well and look forward to the arrival of that day together.
END
Interview | Huan Shi Editor | Xiao Qie Review | Huan Shi