Recently, the Institute of Microelectronics of the Chinese Academy of Sciences issued a stark warning during an internal strategic seminar. Titled The Chasm Between Foundational Semiconductor Research and Industrial Security, the report sharply points out that while China is vigorously catching up in chip manufacturing capacity and applied technologies, there exists a startling "structural deficiency" in foundational semiconductor research and materials science, which will determine the outcome of the next decade. This situation causes the entire industry to advance like entering an unknown "dark forest," facing disorientation and systemic risks.
01 "Barren Foundational Land" Beneath "Application Prosperity": The True Picture of a Mirage and a Capability Gap
Superficial data paints a picture of a booming Chinese semiconductor industry: rapid expansion of wafer fab capacity, the world's largest number of chip design companies, and notable achievements in application-specific chips like 5G communication and image processing. However, this CAS report depicts a starkly different "underlying reality."
The report cites data indicating that over the past five years, China's share of top-tier publications and core patents in foundational fields such as semiconductor materials, device physics, and underlying EDA algorithms has remained below 5% of the global total. This stands in sharp contrast to China's share of over 30% in publications focused on applied areas like process integration and circuit design optimization.
"Most of our research involves engineering optimizations within a known 'map of physics,'" stated the report's lead and director of a key CAS laboratory. "But as the industry approaches the physical limits of Moore's Law, the exploration of next-generation technology paths—such as CFET, atomic channel transistors, and non-von Neumann architectures—depends entirely on a fundamental understanding of new physical effects and new material systems. On this 'new continent' that will determine the future, we barely have the capability to draw the map."
The direct consequence of this "barren foundational land" is that the industry exhibits a "reactive followership" rather than "foresighted layout" in response to external technological evolution. A young researcher involved in the report discussions used the recent hot topic of 2D semiconductor materials as an example: "When the international academic community began systematically exploring quantum transport in heterojunctions of transition metal dichalcogenides (TMDs) five years ago, most of our teams were still focused on achieving better strain engineering on silicon substrates. Now that TMDs have become a leading candidate for the post-silicon era, we find our accumulated knowledge on foundational issues like material growth mechanisms and defect physics is nearly blank."
02 "The Dark Forest" Principle: Three Systemic Risks Caused by Lack of Basic Research
The report creatively borrowed the "Dark Forest" theory to elaborate on the chain of risks that may be triggered by the weakness of basic research:
Risk One: The "Suspicion Chain" and "Dimension Reduction Attack" in Technology Routes
When the foundation research is weak, the industry's judgment on the future mainstream technology routes (such as whether chips will continue to pursue miniaturization or shift to core integration? Whether the next-generation storage will bet on MRAM or FRAM?) is based on extremely incomplete information. It is like trying to judge others' intentions in a dark forest, which makes it very easy to fall into a "suspicion chain" - because one cannot predict the feasibility and ceiling of various routes from the underlying principles, one can only passively follow the hotspots, resulting in resource dispersion and strategic wavering. More dangerously, once the opponent opens up a completely new path from the fundamental principles (such as how FinFET overturned planar transistors in the past), we may face a "principle-level dimension reduction attack", and all the past process accumulations may instantly lose their value.
Risk Two: "Gaps" and "Root Loss" in the Talent Chain
Currently, a large number of outstanding talents from universities and research institutions have been drawn to the "chip design" and "process integration" fields, which can quickly produce results and attract financing. The report warns that this utilitarian orientation is causing a severe generation gap among "scientist-type" engineers who understand the essence of physics, chemistry, and mathematics. An academician expressed his deep concern during the discussion: "The majority of the engineers we are currently cultivating are 'operation manual engineers'. They can proficiently use foreign EDA tools and process design kits (PDK), but few people can deeply explore why the tools are designed this way and when the models in the kits fail at the physical boundaries. Once the 'operation manual' is withdrawn or becomes invalid, we will face the crisis of 'root loss' in the talent system."
Risk Three: "False Safety" and "Islands of Ecology" in the Supply Chain
Although the domestic production rates in areas such as silicon wafers, chemicals, and target materials have been continuously increasing, the report indicates that many domestic materials are merely "copies" without a clear understanding of their underlying principles. "We can replicate the formula of a certain generation of products, but we lack sufficient research on fundamental issues such as how the impurities affect the reliability of the devices and the reaction mechanisms under high temperatures." This means that when international leaders advance to the next generation of products, our supply chain may again become disconnected due to an inability to understand the underlying material requirements. Moreover, the lack of understanding of basic physics also makes it difficult for us to build a deeply integrated device-material-process collaborative innovation ecosystem like that of Intel-ASML, which is based on a common physical understanding. This can easily lead to becoming a technological island.
03 Breaking Out of the "Dark Forest": The Chinese Academy of Sciences Report Calls for the Establishment of a "Principle Innovation" Community
In the face of the severe situation, the report not only issues warnings but also calls for the initiation of a national research system reconfiguration centered on "principle innovation":
Direction 1: Initiate the "Semiconductor Foundation Principles" National Laboratory Cluster
The report suggests that by drawing on the model of the "Semiconductor Research Consortium" (SRC) in the United States, integrate the cutting-edge forces from the Chinese Academy of Sciences, top universities, and leading enterprises, and focus on 3-5 "unexplored areas" that may trigger paradigm shifts (such as the application of topological insulators in interconnection, prototype devices inspired by the brain, and exotic semiconductor materials for quantum computing), establish long-term and stable "seed funds", allowing researchers to conduct high-risk and long-term explorations, with the goal of generating original theories that can be included in future global semiconductor textbooks.
Direction 2: Implement the "Pasteur Quadrant" research paradigm
The report emphasizes that the dichotomy of "pure basic research" and "pure applied development" must be abandoned. A vigorous promotion of applied basic research (i.e., the "Pasteur Quadrant") should be carried out. For instance, major basic research projects should be established around industrial real problems such as the physics of plasma light sources in EUV lithography and the quantum transport of interface states in advanced packaging. A mixed team composed of physicists, chemists, and engineers should be formed to fundamentally solve the bottlenecks.
Direction 3: Reform the evaluation system to support "dark knowledge" explorers
The most crucial aspect is to reform the short-sighted evaluation system that focuses solely on "the quantity of papers" and "project funding". The report suggests establishing a long-term evaluation mechanism centered on "academic influence" (such as whether new theoretical models have been proposed and whether invited to give basic tutorials at top international industry conferences) and "principle-level patents" for teams engaged in semiconductor basic research. This will protect those who are willing to sit in the cold seat and explore the unknown laws in the "dark forest", the "pioneers".
