China's Strategy of Industrial Abundance: an Imperial Banquet, not Molecular Gastronomy

"If a skilled chef already has all the ingredients—vegetables, oil, salt, soy sauce and vinegar—for preparing an imperial banquet, would he really bother with molecular gastronomy?" — Sun Xi (孙喜)

Forming part of what has been described by Dan Wang in his recent book as an “engineering state”, the Chinese policy establishment often lets the term “system” do the conceptual heavy lifting: building a complete “industrial system”, developing a “national innovation system” and strengthening “systemic capacity”. In an increasingly orthodox narrative that is perhaps best exemplified by the economist Lu Feng’s (路风) concept of the “complete industrial system”, technological innovation is understood not primarily as the product of nimble minds or freewheeling vision, but as the natural outgrowth of a holistic system in which supply- and demand-side inputs are calibrated for its emergence.

Echoing the previously covered analysis by public policy specialist Huang Ping (黄平) that China’s advantage in AI derives from its vast “systemic capacity” and pool of industrial demand, management theorist Sun Xi (孙喜) presents a lucid exposition here of how such an industrial system operates—theoretically and in practice. As an overall strategy, he argues that funnelling state funds into research-led frontal assaults on technology “chokeholds” is less effective than strengthening the role of genuine, domestically generated industrial demand in directing research.

An emphasis on demand-side reform might, at first glance, appear like a call for market liberalisation. It is not. The operative principle is “integration”, a term also liberally sprinkled throughout the 15th Five-Year Plan documents, between the goals of the state, industry and researchers.

While enterprises ought to act more as “agenda-setters” for research, they are to do so in a system where “national team” key purchasers align the demands of industry with those of state and society. Though researchers should direct their gaze to the demands of enterprises, the incentive structure proposed for this is highly coordinated: industrial adoption of their research is to be evaluated as part of their performance reviews. Instead of prioritising theoretical breakthroughs by exceptional scientists, the ideal scenario is coordinating researchers’ work within a complete industrial system developed by “politicians, entrepreneurs and engineers”—or, in Sun’s culinary metaphor, opting for an “imperial banquet” over “molecular gastronomy”.

The fine-tuning of such a system, structured around carefully calibrated inputs and perfectly lubricated incentives, evokes the strain implied by economist Yao Yang’s (姚洋) powerful image of “tightening the screws step by step”. As Sun notes with satisfaction, extracting more from this industrial system appears set to be a leading priority during the 15th Five-Year Plan.

— James Farquharson

Key Points

1. China’s prior emphasis on state-coordinated research projects to address technology “chokeholds” risks falling into the trap of identifying research priorities based on US restrictions rather than genuine need.

2. In a new industrial model that closely integrates research with industry, firms should play a stronger agenda-setting role: business history shows that industry–research collaboration outperforms isolated research efforts.

3. The state, as the central coordinator and resource mobiliser, should address institutional barriers and “trust bottlenecks”—for example, ensuring that industrial data from traditional firms is shared with AI software developers.

4. In industrial-use chipmaking and AI software, the high degree of precision and specialisation requires large quantities of sector-specific training data and vertical integration of design, production and application.

5. Hence, compared with other countries’ industrial systems, China’s deep industrial base and sector experience provide a distinct advantage in providing demand-side feedback for industrial-use chipmaking and AI.

I will be in Beijing for the next few weeks, with a short stay in Shanghai. If you would like to meet — whether to exchange views or simply say hello — please feel very welcome to get in touch. — Thomas

6. By contrast, for general-purpose chips and consumer AI, the demand structure is more varied and less specialised: this allows a broader range of internet and IT firms to participate in a horizontal, market-driven model.

7. In recent years, university research priorities have drifted towards award-driven metrics, devaluing “useful science” and weakening the link between research output and real-world demand.

8. Scientific research should therefore be coordinated by politicians, entrepreneurs and engineers, who usually possess a more profound strategic understanding of priorities than the scientists.

9. To strengthen research–industry collaboration, incentives for researchers should reward the successful industrial adoption of their technologies as well as theoretical advances.

10. “Strategic scientists” such as Qian Xuesen—the “father of Chinese rocketry”—are a rare breed; his counsel to prioritise missile over aircraft development drew on an engineering-based understanding of China’s industrial foundations.

Name: Sun Xi (孙喜)
Year of Birth: September 1982 (Age: 43)
Position: Associate Professor and Deputy Director, Department of Enterprise Management, College of Business Administration, Capital University of Economics and Business
Research Focus: Industrial policy; science & technology policy; innovation studies
Other: External Expert, Machinery Think Tank, Machinery Industry Research Institute
Education: BA (2005) and MA (2008) Shandong University; PhD in Management, Graduate University, Chinese Academy of Sciences (2011); Postdoc, School of Government, Peking University (2011-2013)

SUN XI: STOP DEFINING “BOTTLENECKS” BY FOLLOWING THE WEST’S LEAD—THAT ONLY LEAVES US PASSIVE AND BEATEN
Sun Xi (孙喜）
Published by Guancha.cn on 27 October 2025
Translated by Jan Brughmans
(Illustration by OpenAI’s DALL·E 3)

N.B. For the sake of concision, the interviewer’s questions have been summarised.

Question 1: On the pitfalls of an industrial policy model organised around specific projects to address “chokeholds” and the shift to a more demand-driven planning system.

Sun Xi: In recent years, amid profound shifts in the international environment, China’s innovation system has undergone a series of adaptive adjustments [因应性调整]. Two prominent adjustments stand out. The first is the growing emphasis on enterprises as the principal agents of scientific and technological innovation. The second is the strengthening of national strategic scientific and technological capacity in service of overarching national objectives. The crucial nexus linking these two adjustments lies in transforming specific “chokeholds”, encountered by enterprises in the course of innovation and industrial upgrading, into national scientific and technological tasks [任务] and projects [项目].

For instance, massive investments have been directed towards specific research institutions to tackle challenges in photoresist materials and lithography equipment. Similarly, a number of sectoral institutions [行业性机构] have been established to “storm the pass” [攻关] in common technologies across industries, thereby helping the majority of enterprises—particularly small and medium-sized firms—bridge the gaps in their technological upgrading processes [技术升级].

In recent years, this habitual tendency has led many to believe that the main method for the government to assist and guide enterprises in realising their innovative potential is to convert the difficulties encountered by enterprises and industries into specific projects—whenever such challenges arise.

This approach has its merits, yet its function is primarily remedial [事后补救] in nature. By “remedial”, I mean that this institutional mechanism introduces an artificial time lag. Inevitably, several years will elapse from the moment an enterprise genuinely encounters a “chokehold” problem, to the point when the issue is transformed into a national project (proposal submission) and subsequently generates a project outcome (evaluation and delivery). This is because China’s national science and technology projects are planned and implemented on an annual basis.

In the long run, excessive reliance on this approach is bound to give rise to certain problems—for instance, misjudging the extent to which the existing industrial base can support the resolution of “chokehold” issues. In many cases, it is not the absence of technological breakthroughs, but rather specific non-technical factors—social factors such as poor communication between the supply and demand sides—that prevent these achievements from being effectively utilised at particular points in time.

One of the most illustrative historical lessons concerns the breakthrough in developing a high-pressure common-rail system for diesel engines by the Wuxi Fuel Injection Equipment Research Institute [无锡油泵油嘴研究所]; although the product was ready and viable, none of our domestic diesel engine manufacturers adopted it.

At that time, they placed greater trust in the technologies of Germany’s Bosch and Japan’s Denso. This lack of confidence in domestic innovation was later dubbed the “latecomer’s disadvantage” [“后来者劣势”] by Professor Gao Xudong (高旭东) of Tsinghua University. If we neglect this aspect of the issue, we are likely to overestimate the role of government science and technology programmes in driving innovation-oriented transformation and upgrading, or fall into a generalised thought process [泛化的思路].

In other words, regarding our subject today—advancing industrial innovation and transformation through the integration of a proactive government [有为政府] with a well-functioning market [有效市场]—our foremost tasks are the following: to properly design and regulate a unified national market [把全国统一大市场设计好、规范好], make full and effective use of market mechanisms, and mobilise the accumulated cross-sector technological capacities and assorted societal resources of our industrial base. This, in fact, constitutes a crucial dimension of what it means for a government to be “proactive” [“有为”]. In the longer term, this [undertaking] will ultimately be of more significance than project-based models aimed at purely scientific and technological breakthroughs.

Question 2: On empowering enterprises to serve as the principal agents of innovation while closing the gap between their market demands and social/national requirements.

Sun Xi: Only a proper grasp of this question permits a clearer understanding of how the government can effectively assist and guide enterprises in realising their innovative potential. Looking ahead, once we have overcome some of the key “hurdles” [“坎儿”] before us, the government should devote more time and effort to sitting down with leading domestic enterprises, top engineers and strategic scientists to jointly discuss and explore the possible directions of future industrial development.

Through this process of dialogue, the government can not only build its own understanding of industry but also engage in “self-discovery” [“自我发现”] — identifying the key technological nodes and developmental pathways that are crucial to the transformation and upgrading of the industrial system and achieving long-term national goals. This will enable it to define a number of strategically significant future products and projects—as seen in the Xinjiang photovoltaic initiative and the Yarlung Tsangpo River hydropower project [Note: 雅鲁藏布江水电工程; which broke ground in July 2025], which exemplify this mode of thinking. We should no longer fall back on the old intellectual habit of defining our “chokehold” points by following on the heels of the West [跟在西方屁股后面]—a mindset that history and experience have shown only leaves us passive and beaten [被动挨打].

Building on this process of “self-discovery”, the government should further assume the roles of market rule-maker [市场规则设计者]—including setting standards—and key purchaser [关键买家]. By institutionalising these functions, it can continuously raise industry entry thresholds and propel the process of survival of the fittest [优胜劣汰] among enterprises. At the same time, by publicising such initiatives widely, the government can foster a shared long-term vision for all industry participants, break down information barriers between sectors, and guide and mobilise broader participation by indigenous enterprises [民族企业]. Our previous airliner [development] project is a good example: a state-level strategic announcement that drew in participation from a wide range of industries, including machinery, steel, textiles, electronics, software and others. [Note: This refers to a national initiative approved by the State Council in 2007, which led to the creation of the Commercial Aircraft Corporation of China (COMAC) in 2008 and laid the groundwork for the development of the C919 passenger jet; analysts have noted that many of its core components were sourced from overseas.]

The aforementioned measures may serve to bridge effectively the gulf between the genuine market demands [市场需求] of enterprises and the genuine demands of both society [社会需求] and the nation [国家需求]—namely, the so-called “temperature gap” [“温差”]. In essence, a “temperature gap” reflects a deficiency in the state’s capacity for governance and market design.

Question 3: In the context of a policy emphasis on scientific and technological self-strengthening, recognising the need to establish a more distributed planning structure that fosters collaboration between industry, academia and R&D.

Sun Xi: More important [than directly prioritising self-reliance and self-strengthening] is, as in the process I have just mentioned, to further smooth out the collaborative relationships between industry, academia and R&D [捋顺产学研合作关系]. Enterprises that wish to participate ought all to be qualified [有资格] to act as “agenda-setters” [“出题人”] in scientific and technological innovation, thereby enabling a greater number of innovative actors to become fully capable agenda-setters. In this way, a more distributed structure [分布式的结构] can take shape, allowing enterprises encountering complex challenges during the innovation process to be matched in the first instance with suitable partners in universities and research institutes, obtaining prompt resolution rather than becoming mired in national project cycles [通过国家项目绕一个圈子].

This constitutes the fundamental route to diversifying investment in science and technology, as well as an important means of addressing the distortions [异化] that have arisen in national research programmes in recent years. Put differently, in this framework, national projects should no longer serve as “timely rescues” [雪中送炭] but instead as efforts that “enhance value” [锦上添花].

To give an example, in recent years Chairman Xi Jinping has repeatedly noted that China has achieved major breakthroughs in the field of high-end medical imaging equipment. How did these breakthroughs come about? Their origins can be traced back to 2008, when Mindray Medical [迈瑞医疗] approached the Shenzhen Institute of Advanced Technology [深圳先进技术研究院] under the Chinese Academy of Sciences, seeking to jointly develop China’s own high-end colour Doppler ultrasound systems. At that time, global high-end ultrasound technology was monopolised by three companies—Philips, General Electric and Siemens—and the development of high-end ultrasound equipment had become a matter of life and death [生死之战] for China’s entire medical device industry. A young researcher at the Shenzhen Institute, Zheng Hairong (郑海荣)—now vice president of Nanjing University—participated in this industry–academia–R&D collaboration [产学研合作项目] and later emerged as one of the leading figures in the field. As a result, Mindray achieved a remarkable turnaround in the market.

During the same period, United Imaging Healthcare [联影医疗], which later developed domestic MRI equipment, started working with the Shenzhen Institute of Advanced Technology—though it ultimately chose Shanghai over Shenzhen as its base of operations. In this process, the national research and instrumentation projects on high-end medical imaging equipment undertaken by Zheng Hairong and the Shenzhen Institute generally lagged, to varying degrees, behind similar industry–academia–R&D collaboration projects. This exemplifies the role of enterprises as the principal actors in scientific and technological innovation—and the complementary, value-enhancing function of national science and technology projects.

Question 4: On the continuing relevance of vertical integration strategies in certain sectors where market demand is not sufficiently developed—such as industrial-use chips.

Sun Xi: When it comes to high-end chips and industrial software—the so-called “chokehold” areas of high interdisciplinary complexity—there are, in fact, certain differences across industries. Relatively speaking, chips used in the industrial sector can often be developed through a user-led approach to innovation. Hisense’s [海信] entry into television chips and Gree’s [格力] entry into air-conditioning chips both reflect this logic: precisely because they lacked suitable domestic suppliers, they were compelled to address these intrinsic needs through vertical integration.

By contrast, the general-purpose computing chip sector is an industry with a much larger market size, but where the degree of division of labour [分工程度] is far greater. The relationship between pure design firms (such as HiSilicon and other fabless companies) and pure foundries (like SMIC) has already moved beyond a purely transactional nature and is increasingly shaped by external factors—including institutions and policy. Consequently, this sector has already moved beyond the scope of vertical integration as a viable framework.

In contrast, the industrial software sector resembles industrial chips—it is more fragmented [碎片化], has relatively lower demand, and each piece of software embodies the critical technological expertise of the leading enterprises in [the demand-side] industry. Consequently, in recent years, the development of China’s industrial software sector has followed two distinctly visible paths. The first, exemplified by China National Petroleum’s Bureau of Geophysical Prospecting (BGP) [中国石油东方物探], has addressed the software development of key technologies through vertical integration. The second, emerging more recently under the guidance of the Commission for Science, Technology and Industry for National Defence [国防科工委] and coordinated by the Military–Industrial Association [Note: 军工协会 or, in full, the China Association of Peaceful Use of Military Industrial Technology 中国和平利用军工技术协会], has focused on facilitating coordination and transactions between the supply and demand sides of the industrial software sector.

Put simply, both of these paths have moved beyond purely market-based transactional relationships. Rather, they tackle technological barriers [技术壁垒] and trust bottlenecks [信任瓶颈] through more organised and socially coordinated mechanisms.

Question 5: On the differing demand structures for industrial-use and general-use AI and China’s comparative advantage in industrial-use AI.

Sun Xi: A few days ago, I was speaking with colleagues from Zhejiang, who told me that the province is now actively promoting the broad application of artificial intelligence across its industrial sector. Upon closer comparison, we found that the standout performer in this field is SUPCON [浙大中控], a company that originated from Zhejiang University. The reason is clear: over the past thirty years since its founder, Professor Chu Jian [褚健], established the firm, SUPCON has built up exceptionally rich experience in industrial automation and has now become the undisputed leader in process automation control in China.

This has allowed SUPCON to make maximum use of “proof by exhaustion” [“穷举法”]—that is, training artificial intelligence step by step using the structured empirical knowledge accumulated from industrial scenarios. In doing so, it ensures that, rather than relying on the unsupervised inferences of large models, the system develops a basis in clear “if A, then B” judgements, thereby refining its AI solutions and industrial operating system.

More importantly, given the current stage of artificial intelligence development, the exhaustive method remains the single most important means of applying AI to the management of industrial settings—there is no alternative. By contrast, some enterprises originating from the internet or IT sectors find it much harder to apply AI in industrial contexts in the way SUPCON does, as they lack the industrial experience required to implement this “exhaustive method”.

This case actually highlights some fundamental principles for understanding how Chinese enterprises should engage with the evolving landscape of artificial intelligence.

The first fundamental principle is that the application of artificial intelligence varies entirely across different industries. In B2G [Note: “business-to-government”] and B2C [Note: “business-to-consumer”] sectors, the requirements for safety and consistency are comparatively low, and the variables within operational scenarios are comparatively limited. In such cases, large-model technologies—once appropriately adapted—can be used to address certain industry-specific challenges. This is precisely why enterprises originating from the internet and IT sectors have been able to participate deeply in these areas and achieve notable success. Products such as cars, household appliances and personal computers have been among the first to incorporate large models for this very reason.

By contrast, in B2B [Note: business-to-business] industries, the requirements for consistency and safety—as well as the diversity and complexity of variables within operational scenarios—are far higher than in consumer goods or the public utilities sectors. Success in B2B depends heavily on industrial enterprises’ accumulated experience in simplifying and breaking down these [operation] scenarios—such as by developing sets of highly specialised physics-based models [专用的机理模型]—to a level upon which artificial intelligence can comprehend and operate. As such, the industrial experience that Chinese enterprises have built up over the past three decades through participation in global value chains is by no means a burden; it is the “golden key” [“金钥匙”] to China’s engagement with the AI technological revolution. Without this key, we cannot unlock the door to AI-driven industrial development—neither in design, development, production nor marketing.

The second fundamental principle concerns how we view the relationship between traditional industrial companies and artificial intelligence companies. The operational scenarios and experiential data originate from the accumulated knowledge of traditional industrial firms, while the role of AI companies (or certain academic researchers that excel in AI development) is to help traditional industrial enterprises transform that data into mechanistic models and artificial intelligence systems. If this principal–auxiliary relationship [主客关系] is not properly managed, numerous problems may arise—such as the poor communication between supply and demand sides mentioned earlier, which in severe cases can lead to a complete impasse [“此路不通”].

At this stage, some degree of government involvement may be required—for instance, by launching pilot projects in industrial artificial intelligence that provide appropriate opportunities and platforms for collaboration between the supply and demand sides, thereby helping to overcome trust bottlenecks. Ultimately, developing new productive forces [新质生产力] through artificial intelligence requires a concerted effort to effectively manage the relationship between the “new” and the “old”.

Question 6: On Xi Jinping’s understanding of innovation and the importance of strengthening the vital link between industrial demand and innovative research.

Sun Xi: We must first establish a fundamental judgement [判断]: as practice [实践] has advanced, the central leadership’s understanding of basic research has become increasingly deep and pragmatic. We can clearly see this in On Achieving Scientific and Technological Self-Reliance and Self-Strength [Note: 《论科技自立自强》; published in Qiushi in 2023], which compiles General Secretary Xi Jinping’s three speeches to the Two Academies [Note: 两院, here referring to the Chinese Academy of Sciences (CAS) and the Chinese Academy of Engineering (CAE)] in 2014, 2018, and 2021, as well as his 2020 address at the Scientists’ Symposium [科学家座谈会]. In these speeches, Xi’s discussions of basic research—its goal-oriented and demand-driven nature—and of the relationship between basic and applied research fully demonstrate this deepening and increasingly practical understanding.

Looking at developments over the past two years, whether it is General Secretary Xi Jinping’s speech on 21 February 2023 at the Politburo meeting on the issue of basic research [中央政治局关于基础研究问题的讲话], his address on 24 June this year at the Conference of the Two Academies [两院院士大会], or the current 15th Five-Year Plan itself, all of them reflect this same pragmatic character.

Over the past two years, in discussions concerning how to “enhance the overall effectiveness of the national innovation system” [“提高国家创新体系整体效能”], there has been a significant shift in the central leadership’s discourse: an increasing emphasis on the deep integration of scientific and technological innovation with industrial innovation [科技创新与产业创新的深度融合]. At this year’s Conference of the Two Academies, even the long-standing phrase “commercialisation of scientific and technological achievements” [“科技成果转化”] was revised to “commercialisation and application of scientific and technological achievements” [“科技成果转化应用”]. All of these changes reflect an effort within our traditional policy language system [传统的政策语言体系] to align more closely with a new policy paradigm that recognises enterprises as the principal actors in scientific and technological innovation [“企业科技创新主体地位”]. This is highly significant in understanding how we can better bridge the divide between basic research and the practical needs of industry in the new era.

For example, the 15th Five-Year Plan calls for “strengthening the strategic, forward-looking and systemic layout of basic research” [“加强基础研究战略性、前瞻性、体系性布局”], which in fact represents a continuation of the spirit of the 21 February 2023 Speech [on basic research]. A key message of that speech was that basic research should “walk on two legs” [“两条腿走路”], meaning to combine goal orientation [目标导向] with free exploration [自由探索]. The current emphasis on “strategic, forward-looking and systemic” development is essentially a further deepening of the aforementioned goal-oriented approach.

Question 7: On rendering the work of researchers more responsive to strategic and industrial demands.

Sun Xi: Looking further ahead, the effectiveness of a strategic, forward-looking and systemic layout can be maximised only when officials, industry, academia, and R&D [官产学研] come together—as we discussed earlier—to form a shared understanding of the developmental direction of individual industries and of the modern industrial system as a whole. At that point, the state, which is able to steer the overall trajectory of socio-economic development, plays the role of the “agenda-setter” [“出题人”] for basic research. Moreover, only when the state has a clear grasp of that developmental direction, can such “agenda-setting” become operable—and, to the greatest extent, useful. Yet, judging from the current evaluation methods for “organised research” [“有组织科研”] in universities, our understanding of this issue still has room for considerable improvement.

Of course, innovation is inherently uncertain, which means it is impossible to “bet correctly” [“押中”] on all major basic research topics one or two decades in advance. The state, therefore, cannot be the sole “agenda-setter”. At this stage, we must develop a deeper understanding of companies’ role as the principal agents of scientific and technological innovation. For example, as I mentioned earlier, the Shenzhen Institute of Advanced Technology (SIAT) [深圳先进技术研究院] under the Chinese Academy of Sciences is a typical representative of the new type of research institutions established in the early 21st century: positioned as an industrial research institute with the resources of a national team [Note: 国家队; refers to state-backed institutions or enterprises], and—crucially—rooted in the industrial soil of Shenzhen and the Pearl River Delta.

My observation of the Shenzhen Institute of Advanced Technology (SIAT) began in 2018, and I have summarised its practice of promoting the deep integration of scientific and industrial innovation as the “big tree model” [“大树模型”]. In this model, scientific innovation and basic research within the institutes form the roots; industrial development is the canopy; the new products created by enterprises are the fruit; and deep integration relies on the trunk. The trunk’s role is not only to carry the water and minerals absorbed by the roots up to the canopy, but—more importantly—to transport the organic compounds produced in the canopy back down to the roots, thereby providing them with the energy needed for growth.

At SIAT, researchers are willing to accept the energy transmitted from the canopy (the industrial side)—that is, demand-side knowledge and funding—because their performance review [考核机制] explicitly includes a section on adoption by industry. Publishing papers alone is not enough; you must also understand the language of industry, be able to undertake horizontal projects and translate industrial needs into scientific questions—in short, pursue “useful science” [“有用的科学”].

Since goal-oriented [目标导向] basic research can take shape through both top-down and bottom-up approaches [“自上而下”和“自下而上”], we should cautiously evaluate the new tendencies that have emerged amid recent reforms in the scientific and technological system. For example, if research is expected to “serve major national strategic needs” [“服务国家重大战略需求”], at what scale [尺度] should the notion of “major” be defined and assessed? Although the pursuit of publication in English-language journals has yet to fade completely, it remains essential for the country’s strategic science and technology institutions to focus on their primary missions and concentrate their efforts where they can have the greatest impact.

However, if the Chinese Academy of Sciences were to delegate the authority to define one’s “core responsibilities and main business” entirely to its subordinate branches or specialist institutes, a narrow interpretation of these roles could come into conflict with the historically established inter-disciplinary structures of each institute. In extreme cases, this might even lead to the problem where, as the saying goes, “you can’t plant beans in an aubergine plot” [“茄子地里不能种豆角”]—in other words, research organisations operating in a specifically defined field (A) may be unable to work on projects for research group (B) that are geared towards national strategic needs.

Another example is collaborative innovation between research institutions and leading industrial companies, particularly in laying out seed projects in frontier fields. Can the amount of funding truly reflect the projects’ major significance [重大性] and strategic value [战略性]? If research institutions single-mindedly chase large, headline projects or approach industry–academia collaboration with the mentality of “charging towards performance targets” [“冲业绩”]—while shying away from smaller initiatives with long-term importance—they risk missing the genuine from 0-to-1 breakthroughs in the evolution of technology and industry. These problems, which have become evident in practice, urgently require greater attention and effort from policymakers.

At the other end, under the influence of disciplinary evaluation mechanisms [学科评估机制], the university system over the past two decades has steadily—or at least for a time—drifted away from industrial needs and persistently devalued horizontal projects [Note: involving collaboration with industry rather than reliance on centrally-disbursed funds], coming to interpret research and basic science in an increasingly narrow and insular manner. Therefore, the most urgent task in addressing the disconnect between basic research and industrial demand is to reshape the research system—especially universities’ capacity for, and imagination about, demand orientation [需求导向]—removing constraints and broadening horizons.

Only by equally weighing every penny from vertical [Note: government-funded] and horizontal [Note: industry-commissioned] projects can universities and research institutes rebuild their enthusiasm for serving the needs of industrial innovation. In the long term, we must rethink and reconstruct the public discourse surrounding basic research. For instance, excessive emphasis on the so-called “usefulness of the useless” [“无用之用”] in basic research—even to the point of denying the very possibility of “useful science” [“有用的科学”]—or condemning any discussion of the usefulness or demand orientation of basic research as “utilitarian” [“功利”] is misguided. After all, has the utilitarian tinge of the scientific system not long been the result of evaluation criteria and remuneration structures overly dependent on research awards [科研奖励]? And, for quite some time, have these evaluation criteria truly been demand- or goal-oriented?

In fact, the “walk on two legs” [“两条腿走路”] formulation in Xi Jinping’s 21 February 2023 Speech [Note: as per above, combining “goal-orientation” with “free exploration”] already offers a profound answer to this question. Rebuilding the public discourse on basic research will also help, at the societal level, to reshape the relationship between scientific and industrial innovation—especially since many comrades still hold rather rose-tinted views of what scientific innovation entails.

I once heard a comrade from Hebei, at a meeting on coordinated industrial development in the Beijing–Tianjin–Hebei region, engage in a self-criticism [自我批评] where he said that the poor conversion of scientific and technological achievements in the province was due to its weak industrial base, which made it difficult to find application scenarios for Beijing’s innovations locally. I tried to reassure him, saying, “You’re being far too modest! In fact, even many of Beijing’s own scientific innovations can’t find practical application.” From an institutional perspective, however, we must recognise that such “humility” [from the industrial application side] is not the right attitude or mindset for promoting deep integration [of basic research with industrial applications]. From this standpoint, we need to demystify [祛魅] basic research.

Question 8: On engineering an industrial system that leads to innovation through practice, rather than isolated theoretical breakthroughs.

Sun Xi: This is quite an abstract question. To illustrate it to everyone more clearly, let me tell two stories.

The first story is one I particularly like to tell—the story of China’s breast milk research. A large part of this research concerns the composition of breast milk and the functions of its various components. Although China began breast milk research quite early, for a long time it remained fragmented [很零散] and unsystematic [不系统]. Systematic, long-term research on breast milk only emerged in China when a group of dairy enterprises—represented by Sanyuan [三元], Feihe [飞鹤], and Yili [伊利]—began pursuing in-house innovation [自主创新].

For example, Sanyuan established a breast milk database containing tens of millions of data points and conducted long-term studies on breast milk composition, microbiota (probiotics), the functions of various components and their preparation processes. This work laid the foundation for Sanyuan’s in-house developed infant formula with Chinese characteristics [中国特色的配方奶粉] and a range of innovative fermented yoghurts. As a result, Sanyuan’s chief scientist, Chen Lijun (陈历俊), became affectionately known as “the grandfather of Chinese babies” [中国宝宝的姥爷]—both because he has a daughter of his own and because he is, in fact, a grandfather. More importantly, Chen remains the most prolific scholar in the field of breast milk research in China. This serves as a quintessential example of industrial development driving technological advancement, and technological advancement in turn propelling scientific research.

The second story comes from the agricultural sector in Anhui Province. The company involved is called Geyi [格义]. Although Geyi has faced difficulties for non-technical reasons, it developed a particularly intriguing product. During deep processing of crop straw [秸秆], the company produced a liquid by-product and unexpectedly discovered that it could increase grain yields and restore soil fertility—specifically, by rehabilitating saline-alkali soil.

Upon further investigation, they found that the liquid’s effectiveness lay in its active components, which not only could sequester [Note: 螯合, to chelate] large quantities of heavy metal ions in the soil, thereby alleviating the salt stress that saline-alkali soil imposes on plant seeds and roots. It could also provide ample nutrients for the growth of beneficial microorganisms, thereby enhancing crops’ resilience and enabling them to better withstand adverse growing conditions.

In Geyi’s own words, “Every experienced farmer understands the principle, but only we have the material.” In other words, the scientific principle had long been known—but no one knew how to put it into practice or what to use to achieve it. Why was that? Because Geyi’s deep-processing technique for crop straw was developed entirely in-house [自主开发], the discovery of this liquid was therefore an original one.

Even more interestingly, when Geyi sought to promote this revolutionary product, it encountered an awkward regulatory constraint: under Chinese law, fertiliser products are required to contain nitrogen, phosphorus and potassium. Yet their product was an organic solution that included none of these traditional fertiliser elements. As a result, they were compelled to add a certain amount of nitrogen, phosphorus and potassium in order to obtain the market entry approval [“准入证”] for the fertiliser market.

From these two examples, we can easily discern that the defining feature of innovation is its practical nature. Put simply, innovation is achieved through doing—not through mere speculation or calculation. Wherever practice advances, innovation can advance; and wherever innovation advances, so too can “useful science”.

Thus, “useful science” has national boundaries [国界]—it is defined and shaped by a country’s own industrial system [工业体系].

The situation in China is unique: we possess an industrial system of a scale and diversity unprecedented in world history, providing an exceptionally rich foundation for industrial innovation. This abundance of practical experience and resources has, in turn, greatly lowered the barriers to major innovation—both in creation [做出来] and commercialisation [市场化].

Thus, by fostering the creative spirit of the people and encouraging every sector to take an active role in innovation, China can revitalise its entire industrial system—a process that will, in itself, generate more and greater breakthroughs [重大创新]. The 15th Five-Year Plan reflects this logic by prioritising “building a modernised industrial system” [“建设现代化产业体系”] ahead of “achieving high-level scientific and technological self-reliance” [“高水平科技自立自强”], designating it as the first of the plan’s twelve key focus areas.

However, we must be cautious about the topic of the so-called “strategic scientist” [Note: “战略科学家” —a term in Chinese policy discourse referring to scientists that align their research with national strategic objectives]. Returning to the aforementioned policy paradigm of government–industry–academia–R&D coordination [官产学研协同的政策范式], it transpires that the strategic decisions defining an industry’s future direction are not made by “scientists” alone. Mastery of national strategy, insight into competitive dynamics, and understanding of industrial foundations rely far more on the strategic vision [战略眼光] of politicians, entrepreneurs and engineers.

When Mr Qian Xuesen [Note: 钱学森 (1911-2009), best known as the father of China’s space programme and rocketry] is invoked as the quintessential example of a “strategic scientist”, many overlook the fact that “missiles first, aircraft later” [“先导弹后飞机”], his most significant recommendation for the Twelve-Year Plan [Note: “十二年科学技术规划”, the Twelve-Year Scientific and Technological Plan], stemmed from a profound understanding of China’s industrial foundations. They likewise forget that the work Engineering Cybernetics [《工程控制论》], a key expression of Qian’s intellectual legacy, was in its purpose, content and influence firmly grounded in the realm of engineering.

A clear understanding of the practical nature of innovation also sheds light on why, in today’s discussions of “original innovation”, Western countries—particularly the United States—show such a strong preference for major theoretical breakthroughs [重大原理突破] that yield fundamental discoveries. Deindustrialisation has eroded their practical foundations for industrial innovation, forcing them to pin greater hopes on new theoretical models that rely less on tangible, practice-based material.

But let us pose a rhetorical question: if a skilled chef already has all the ingredients—vegetables, oil, salt, soy sauce and vinegar—for preparing an imperial banquet, would he really bother with molecular gastronomy?

READ MORE

---

---