For much of the past four decades, the transformation of China’s cities was most visible in their physical form. New districts appeared, industrial parks expanded, transport networks spread, and millions of people moved into an urban economy being built at extraordinary speed. To many observers, construction itself became the defining story of Chinese urbanization.
Yet the physical city was only the visible expression of a broader economic development model. Land-related revenues financed infrastructure; infrastructure attracted companies, investment and people; industrialization expanded the tax base and created employment; and the resulting growth enabled cities to invest again. This self-reinforcing relationship between land, infrastructure, industry and urbanization was well suited to a country moving rapidly from an agricultural to an industrial economy.
Today, the conditions that sustained that model are changing. Land cannot be supplied indefinitely, property cannot remain the principal engine of local growth, and good infrastructure is no longer sufficient to distinguish one ambitious Chinese city from another. The next phase of competition will therefore be determined less by the urban space that cities can create than by the depth and density of the industrial capabilities they can accumulate. After building the physical foundations of development, the central question is how a city can develop the less-visible capabilities that allow innovation to flourish, companies to scale, and industries to renew themselves.
ChangXin Memory Technologies’ (CXMT’s) move toward a public listing on Shanghai’s STAR Market offers a useful starting point. This Hefei-based memory-chip manufacturer has passed the Shanghai Stock Exchange’s listing committee review and received registration approval from the China Securities Regulatory Commission. For this company, this is an important capital-market milestone. The proposed IPO is expected to value CXMT at approximately $350 billion, potentially making it China’s most valuable publicly listed company. For Hefei, however, it represents something larger: An industrial commitment sustained over many years is beginning to mature, linking technology, capital, suppliers, talent and the long-term transformation of the city.
Hefei is not an isolated case. Wuhan has established meaningful scale in memory chips and optoelectronics. Suzhou has combined a deep manufacturing heritage with clusters in optical communications, semiconductors and intelligent manufacturing. Hangzhou has become home to a new generation of technology companies, including DeepSeek, Unitree Robotics and BrainCo. These cities began with different endowments and followed different paths, but they point to the same underlying shift.
In my view, this is one of the most important changes taking place in China’s development model. It is not simply a sectoral rotation from real estate into high technology. It is a shift in the mechanism of growth itself — from expanding physical assets to building systems in which knowledge, institutions, companies and industrial networks reinforce one another.
From projects to ecosystems
Under the earlier model, a local government could prepare land, build roads and utilities, establish an industrial park and offer incentives to attract investment. The logic was largely project-based: secure an investment, build the surrounding infrastructure and use the resulting activity to support another cycle of expansion. This model created enormous productive capacity, even if its limitations have since become more apparent.
The new model is more demanding because industrial capability cannot be assembled through physical investment alone. A semiconductor, robotics or artificial-intelligence company requires not only facilities and funding, but an interconnected base of research institutions, engineers, specialized suppliers, production equipment, patient capital and sophisticated customers. These elements are owned and controlled by different actors. The central task is therefore not merely to provide them, but to connect them so that knowledge can move, products can improve and companies can scale.
Wuhan illustrates how large strategic projects are becoming more central to the way cities deploy resources and define their economic future. Reported cumulative investment in the first three phases of Yangtze Memory Technologies Corp has exceeded 270 billion yuan ($39.9 billion). This figure should not be treated as directly comparable with annual land-related revenue, nor does it represent municipal spending alone. Its significance lies instead in the scale and duration of the industrial commitment — and in what such a commitment can build around it.
An anchor company creates demand for engineers, research, equipment, suppliers, logistics and financing, while attracting customers and complementary businesses. As these capabilities develop, they draw in additional companies and talent, making the location more valuable to the original enterprise. Over time, the process can become self-reinforcing. The company changes the industrial structure of the city, while the city becomes part of the company’s competitive advantage.
An industrial ecosystem is therefore more than a collection of companies operating in the same place. It is a system of learning. Engineers move between firms, suppliers solve problems with customers, manufacturers refine products through repeated production, investors develop sector knowledge, and research institutions gain a better understanding of commercial needs. The real advantage lies not only in the assets a city possesses, but in the speed and quality with which knowledge circulates among them.
Suzhou illustrates this distinction. Its position in AI and hard technology is not explained simply by the emergence of several promising companies; it rests on a broad and mature manufacturing base. In 2025, high-tech industries accounted for 56.2 percent of the output of the city’s major industrial enterprises. For a technology company that must repeatedly prototype, work with suppliers, refine its design and move into scaled production, this industrial depth can shorten the distance between scientific possibility and commercial reality.
Competition in many strategic industries therefore takes place at two levels at once. Companies still compete through technology, products, costs and business models, but behind them sit ecosystems that influence how quickly they learn and how effectively they execute. A strong company can help create an ecosystem; over time, a strong ecosystem can produce and support multiple strong companies. The relevant unit of competition is no longer the company alone, but the company together with the system that shapes its ability to learn and execute.
How direction, experimentation and market selection interact
Why have some Chinese cities been able to organize such ecosystems? Industrial policy and government support matter, but neither alone explains the outcome. China’s development has emerged from a more complex interaction among national direction, local experimentation and market-based entrepreneurial activity.
At the national level, the central government identifies strategic priorities based on technological change, industrial security, development needs and judgments about the sources of future competitiveness. The growth trajectories of electric vehicles, semiconductors, AI and robotics have all been influenced by such judgments. National policy can provide direction, research funding and infrastructure, sending long-term signals to local governments, investors and companies.
But direction is not the same as detailed control. The central government does not determine which companies will succeed or prescribe how every industry must develop. Between national intent and business outcomes lies a wide field of interpretation, experimentation and competition.
Local governments operate in this middle space. A capable city must translate broad direction into a strategy grounded in its own reality: The quality of its universities, the experience embedded in its factories, the availability of specialized talent, the depth of its capital base, the competence of its institutions and its access to markets. Because these conditions differ substantially across China, the same national priority can produce different local approaches. This variation allows the system to explore more than one path.
Companies and entrepreneurs then respond not only to policy signals, but also to customers, technology, costs and competitors. Government can improve the conditions for innovation, but it cannot substitute for entrepreneurial judgment, create a sustainable business model or persuade the market to accept an inferior product.
Seen in this way, China’s three-level structure is not simply a hierarchy through which instructions move downward. Its value lies in the feedback among the levels: national priorities focus attention, local experimentation tests different ways of organizing resources, and market competition reveals which technologies and business models can create sustainable value. Information also moves upward. Successful models are scaled, failures generate learning, and policies evolve as conditions change.
This process is neither automatic nor always efficient. Local experimentation can produce duplication, overinvestment and a rush into fashionable sectors. Policy enthusiasm may run ahead of customer demand, while capital may continue to support companies that are strategically attractive but commercially weak. The value of the system depends not only on its capacity to mobilize resources, but also on its ability to learn, correct and withdraw.
This is why the quality of local execution matters. Local officials stand close enough to companies to understand practical bottlenecks while possessing the authority to assemble capital, infrastructure, institutions and public platforms at scale. When they understand both the industry and the appropriate limits of government intervention, they can translate national priorities into locally relevant strategies without removing the role of market selection. When they do not, support can become protection, and investment can become a substitute for competitiveness.
Hefei and Hangzhou: two paths to ecosystem advantage
Hefei and Hangzhou illustrate two different routes to ecosystem advantage: one built around concentrated commitments to strategic anchors, the other around a broader platform for entrepreneurial emergence.
Hefei has often behaved as an active industrial participant. It has concentrated resources on a limited number of industries capable of reshaping its economy, using state capital, industrial funds, infrastructure and dedicated teams to attract anchor companies and build value chains around them. From BOE Technology and CXMT to NIO, Hefei has moved beyond conventional investment promotion, treating capital deployment, company growth and supplier development as connected parts of a long-term process of industrial formation.
This anchor-led model uses concentrated commitments to create demand for suppliers, talent, research and capital. If the initial choices are sound and execution is sustained, these capabilities accumulate around the anchors and gradually reshape the city’s industrial identity.
Hangzhou has followed a different path. Its role has been closer to that of an ecosystem enabler, creating an environment from which a broader and more diverse base of entrepreneurs and technology companies can emerge. Research investment, industrial funds, talent policies, pilot-production platforms and opportunities to test new applications have all contributed. In 2025, research and development spending among the city’s major industrial enterprises reportedly exceeded 90 billion yuan, while the stated scale of its “3+N” industrial-fund system surpassed 300 billion yuan. The emergence of a group of fast-growing technology companies popularly labeled the “Hangzhou Six Little Dragons”, together with other startups, reflects not only the quality of individual founders but also the density and openness of the system around them.
Hangzhou’s approach is therefore more platform-led. Rather than organizing development mainly around a small number of selected anchors, the city has sought to increase the probability that different kinds of companies can emerge, experiment and grow. The emphasis is less on predicting every winner and more on creating conditions in which winners can be discovered.
Neither model is inherently superior. Anchor-led development may suit capital-intensive industries requiring large initial commitments and coordinated supplier formation. Platform-led development may be more effective where entrepreneurial density, research capacity and market openness already support diverse forms of innovation. The appropriate choice depends on a city’s starting conditions, institutional capabilities and willingness to sustain a commitment long enough for industrial learning to compound.
This also explains why the visible instruments used by successful cities are easy to copy, while the institutional capabilities behind them are difficult to reproduce. Any city can establish an industrial fund, open a science park, introduce a talent program, or announce support for artificial intelligence and robotics. But funds, parks and programs amount to an ecosystem only when they are connected by industry knowledge, disciplined governance, real customer demand and continuous adaptation.
The strategic asset is therefore not the fund, the park or the incentive itself, but the institutional capability to combine resources, learn from the market and remain patient without becoming passive. That capability takes time to develop and cannot be imported through a standard policy package.
The changing competence of local government
This shift is redefining what “competent local government” means. Under the earlier model, competence was demonstrated by assembling land, delivering infrastructure, attracting investment and completing projects. Those capabilities remain relevant, but they are no longer sufficient.
Local governments increasingly need to understand how industries and technologies are evolving, how supply chains are structured, and where companies face constraints. They must be able to distinguish between a shortage of capital and a shortage of customers, between a scientific breakthrough and a scalable product, and between a promising sector and a temporary wave of enthusiasm.
This requires a different type of organization. A local government cannot possess all the knowledge held by entrepreneurs, engineers, investors and customers, but it can create mechanisms through which that knowledge is brought together. Its role is becoming less like that of a land manager and more like that of an ecosystem integrator — bringing together institutions, knowledge and resources without attempting to control every commercial outcome.
However, an integrator is not an operator. Government should not attempt to replace the entrepreneur or manage the company. Its more difficult responsibility is to know when direct participation is necessary, when enabling infrastructure or patient capital is more useful, and when the correct decision is to step back and allow competition to determine the outcome.
This boundary is especially important in semiconductors, AI and robotics, where development cycles can be long and commercial outcomes uncertain. A research breakthrough must be absorbed by a company and translated into a reliable, economical product; manufacturing must reach scale; capital must tolerate uncertainty; and customers must validate the result. If one of these links is missing, a city may accumulate impressive projects without creating a durable industry.
Government can reduce the friction between these stages, but it cannot bypass them or eliminate failure from innovation. One measure of a mature ecosystem is not that every company succeeds, but that resources, knowledge and talent can be redeployed when some do not. Without this capacity for renewal, an ecosystem can become a collection of protected incumbents rather than a source of continuing innovation.
What global companies need to understand
For global business leaders, the emergence of city-based industrial ecosystems requires a more granular way of reading China. China has never been a homogeneous market, but city-level differences are becoming more strategically important as innovation, manufacturing and specialized capabilities concentrate in particular urban systems.
The traditional location checklist — labor costs, tax incentives, land availability and physical infrastructure — captures only part of what now matters. Companies must assess not only cost and infrastructure, but also the depth of local research, production, capital and market capabilities. In many advanced industries, choosing a city increasingly means choosing a learning, innovation and production system.
This changes both competition and collaboration. A multinational may believe it is competing with an individual Chinese company. In reality, that company may draw strength from an entire network of suppliers, engineers, investors, research institutions and early-adopter customers. The same ecosystem can also become a partner: a source of suppliers, research relationships, application environments and forms of experimentation that a multinational could not easily build alone. The question is not simply whether to enter a particular location, but what role the company can play within the local system and what capabilities it can contribute, access or develop through participation.
Global companies may therefore need to supplement a national China strategy with a portfolio of city-level strategic roles. One city may serve as a center for research, another as an advanced-manufacturing hub, another for application development, and another as a source of entrepreneurial partnerships. Managing this portfolio will require local understanding, organizational flexibility and the ability to connect China-based capabilities with the company’s wider global network.
CXMT’s progress toward a public listing is one signal of a much broader transformation. The logic of Chinese urban growth is shifting from the conversion of land into capital to the conversion of knowledge, coordination and industrial relationships into capability. The transition will not be even. Some cities will make poor choices, some industries will become overcrowded, and some local governments will discover that the ability to mobilize resources is not the same as the ability to build competitiveness.
The direction is nevertheless becoming clearer. China’s next phase of urban development will depend less on how extensively a city can expand than on how effectively it can learn, connect, adapt and renew. The cities that develop these capabilities will not simply host successful companies; they will become part of the reason those companies succeed.
For global business leaders, the question is therefore not whether these urban ecosystems matter, but how deeply they understand them and how intelligently they position their companies within them. This is the new map against which strategies in China — and increasingly strategies in relation to China — will have to be formed.
The author is founder and CEO of Gao Feng Advisory Co, a strategy and management consulting firm with roots in China.
The views do not necessarily reflect those of China Daily.
