The curious case of Chinese innovation

A decade ago, the Chinese government launched a comprehensive blueprint for industrial development, ‘Made in China 2025’. This plan both reflected and reinforced the central role of advanced technology and innovation in the next stage of China’s economic development. Since its launch, the plan’s objectives have been expanded and refined in scores of initiatives, regulations, and investment programmes. To realise these, Chinese authorities have deployed a vast array of policy tools, ranging from research funding, subsidies, and tax reductions to infrastructure construction and curriculum reform in higher education. 

‘Made in China’ in 2025

The ‘Made in China’ plan listed ten key industrial sectors in which China intended to gain global leadership: information technology, robotics, green energy and vehicles, aerospace equipment, maritime engineering, railway equipment, power equipment, new materials, pharmaceuticals and medical devices, and agricultural machinery. Ten years later, China has met, or even exceeded, many of those goals. The country is, for instance, now the undisputed global leader in solar panels and electric vehicles, commanding over 80% of the global solar panel market and dominating exports of electric vehicles and related technologies. 

China’s chip industry has also developed far more rapidly than expected, despite multiple rounds of US export restrictions – although it still faces many challenges in directly competing with the most advanced global semiconductors. In artificial intelligence, Chinese researchers have closed much of the gap with their US counterparts. The launch of DeepSeek’s V3 model in late 2024 showcased China’s ability to develop AI models and that Chinese AI researchers can not only develop models performing in the same league as US leaders such as OpenAI and Anthropic, but that those models run far more efficiently, using less power and requiring less advanced chips. 

To be sure, Made in China has not fully achieved its goals. In civil aerospace, for instance, Comac is still a very long way from emulating the success of industry leaders Boeing and Airbus. In other sectors, much remains to be seen: China has made considerable progress in humanoid robots, but these machines still struggle to complete more intensive tasks such as finishing the Beijing half-marathon, especially under adverse conditions like strong winds. Not all investments have been worthwhile: Chinese authorities have spent heavily on the construction of data centres and networks, but many are poorly run or unsuited to the needs of AI development. This has reportedly prompted economic planning authorities to launch a review of the entire sector. Much investment has been spent in wasteful or corrupt ways, with the country’s biggest state-owned semiconductor investment fund, the ‘Big Fund’, a particularly problematic example of corrupt practices. In May, the former chairman of chipmaking conglomerate Tsinghua Unigroup was sentenced to death for embezzlement. 

Perhaps most impactfully, the policy’s focus on self-reliance and technological leadership was itself a response to technological security concerns. The Chinese government wanted to avoid dependence on US technology, fearing espionage (see Snowden), cybersecurity risks (see Microsoft’s discontinuation of security support for Windows XP), and sanctions (semiconductors). But the policies they have implemented have accelerated the very geopolitical tensions and anti-Chinese pushback among Western governments that exacerbate the technology-related pressures that the plan was meant to address.

China’s way

Nevertheless, there are many valuable lessons that can be learned from both the successes and the failures of China’s targeted support for key sectors. In Western discourse, industrial policy often carries negative connotations, conjuring images of Soviet-style planned economies or governments ‘picking winners’, with all the corruption, favouritism, and market distortion that entails. 

Yet China’s approach has often differed, blending broad market mechanisms within a goal-oriented policy framework that guides commercial actors toward national objectives. A first element of this policy is the government’s role as a provider of supporting infrastructure. China’s transition to electric vehicles has been assisted tremendously by local governments actively installing charging stations. The country’s sleek high-speed passenger trains attract the most media attention, but its freight rail system has seen significant growth as well. In order to facilitate the development of the AI sector, pilot projects are underway to create a nationwide ‘compute network’, aiming to provide computational power as a utility similar to water and electricity. 

The second is vertical integration of industrial ecosystems. China’s success in electric vehicles is built on an abundance of raw materials, such as rare earth elements, and knowledge to utilise them, for example, in components such as batteries. Unlike other developing economies, China is less likely to shed the lower value-added parts of this supply chain. This is partly for security reasons: China does not intend to end up with the same dependencies that Europe and the US have come to lament. But it also creates development potential: clustering related activities together can lead to mutual learning and greater efficiency across supply chains. 

The third is intense domestic competition. Many of the ‘Made in China’ key industries are in 2025 highly competitive. In the electric vehicle sector alone, at least 100 brands vie for customers. Inevitably, a wave of consolidation will take place. In the meantime, this fierce competition drives these firms to rapidly iterate and improve their products for the domestic market and prepares them well for international success.

That’s not to say that China’s industrial policy is successful everywhere. It has run into difficulties in certain specific sectors. In aerospace, for instance, Comac faces incredibly high barriers to entry, not only because its products are a generation behind those of its competitors, but also because it must convince airlines of its ability to provide maintenance and support services around the world in the way Airbus and Boeing are able to. In biomedicine, the long development timeline of drugs, treatments, and other innovations means it may be too early to assess the success of efforts in this field. 

Censors and circuits

Even so, the results of Chinese industrial policy disprove the idea, broadly shared not that long ago, that China ‘cannot innovate’, either because of its Confucian heritage or its Communist present, both of which allegedly conspire to rob potential innovators of the freedom their creativity requires. There are, of course, clear political limitations to, for instance, the abilities of Chinese large language models. A routine trope, whenever a new Chinese AI model comes out, is to ask it what happened on Tiananmen Square in June of 1989. Although Chinese large language models may be unable to address such sensitive topics, in many areas of industrial policy, the freedom to create content is simply not important. Semiconductors, power generation equipment, or food delivery robots are in no way affected or constrained by content controls. In other words, the impact of regime type on innovative capacity can be, and usually is, overestimated.