China’s “state-as-platform” model reshapes the global tech race

In its 15th Five-Year Plan, the government has reemphasized emphasized once again developing “new-quality productive forces,” with the ambitious goal of making China a world leader in areas such as artificial intelligence, quantum technology, biotechnology, clean energy, drones, and 6G.

Understanding China’s technological ambitions, however, requires moving beyond the conventional view of its innovation system as one entirely driven by top-down state planning. In many of the most dynamic sectors – from electric vehicles (EVs) and batteries to solar energy and AI – leading companies are privately owned, with limited government involvement. Market competition is often so fierce and margins so small that the government occasionally intervenes to curb what it calls “excessive competition.”

“Government” platform

A more accurate way to describe China’s innovation system is to think of it as a platform organized by the state. Rather than replacing the market, the government structures it – setting strategic direction, establishing pilot zones, facilitating procurement, coordinating standards, and providing regulatory support. Companies then enter these markets and compete intensively, incentivizing rapid improvement and large-scale adoption.

China’s approach resembles that of platform companies such as Nvidia or Apple: the government creates an architecture within which companies compete while regulating interactions in the broader ecosystem. In this model, profitability at the company level is not the primary measure of success. Instead, value is accumulated at the system level through cost reduction, supply chain integration, accelerated learning cycles, and adoption of new technologies, all of which strengthen national capabilities.

This dynamic is already evident in sectors where China has come to dominate the global market. In solar panels and electric vehicles, intense competition has often eroded profit margins while strengthening China’s international competitiveness. BYD illustrates this trend: it recently overtook Tesla to become the world’s the world’s largest electric car manufacturer despite a domestic price war that put pressure on its profitability.

Another example is solar energy. Many companies are struggling financially, but China now controls over 80% of the solar industry. more than 80% of the solar panel supply chain, with intense competition in the domestic market helping to consolidate production and strengthen the country’s global advantage.

AI now appears to be following a similar trajectory. Chinese AI companies, many of which have moved to open source models, are under significant financial pressure. But open source also serves a strategic purpose, accelerating the spread of AI capabilities and closing the gap with leading U.S. labs.

At the same time, the Chinese government is strengthening the ecosystem as a whole. The National Data Bureau has issued guidelines to promote high-quality training datasets, and the Ministry of Industry and Information Technology is pushing forward the creation of a a nationwide computing network to interconnect and standardize architectures, thereby removing capacity constraints and mitigating chip shortages.

Support through pilot programs

As physical AI – including autonomous vehicles, robots, and drones – moves from development to deployment, the next phase may be even more significant. Because these systems require extensive interaction with the real world, data has become a major bottleneck. As one of us (Zhang) argues in a recent article, China’s advantage lies in its ability to create a “data flywheel” in which implementation generates continuous streams of data, which in turn improves performance and reliability.

Government-backed pilot programs play a key role in this effort, accelerating adoption and facilitating data collection. In 2022, Baidu received China’s first approvals for commercial fully unmanned robotaxi services. In three years, Chinese robotaxi companies have expanded their operations by more than to 20 cities across the country, as well as markets in the Middle East and Europe.

Setting standards has also become an important tool to reduce costs and ensure data interoperability. To coordinate the complex supply chains required for humanoid robots, Chinese authorities have brought together more than 120 companies and academic institutions and industry groups to develop common standards.

Just as companies such as Meta and Alphabet rely on proprietary data as a competitive advantage, China increasingly views data generated by physical AI as a strategic national asset. New rules for cross-border transfer of automotive data, finalized earlier this year, impose stricter controls making it harder for foreign companies to use Chinese data to train models. The rules are already having a tangible effect: Tesla’s rollout of its advanced driver-assistance systems has been has been delayed despite the company’s significant presence in the Chinese market.

None of this is to say that the Chinese model of innovation is inherently superior to the Western model. The United States remains more open, more pluralistic, and often more effective at creating technological breakthroughs. China’s approach also comes with serious risks, including misguided investments, chronic overcapacity, and widespread inefficiency.

However, the influence of the Chinese state as a platform should not be underestimated. If the ultimate goal is artificial general intelligence – which will potentially bring enormous benefits to whoever develops it first – the U.S. may well retain the lead. If value is shared through low-cost deployment, industrial integration, and ecosystem-wide learning, America could end up creating the most advanced models and still lose the race to shape the industrial revolution that AI is poised to unleash.