Why China believes it will win the global AI race
State control of financial capital and evolving factor endowments are seen as the keys to eventual victory.
The standard narrative is that post-Mao China saw a rapid rise in wealth on the back of market reforms that dramatically reduced the role of the state in the workings of the economy.
The 1980s and 1990s saw the gradual dismantling of the Soviet-style command economy, the swift introduction of markets powered by private enterprise, and the scaling back of party influence upon management decisions and resource allocation.
A top economist at China’s largest bank, however, has just mounted an argument for the superior efficacy of the Chinese system which runs significantly against the grain of this long-standing narrative.
Cheng Shi (程实) - chief economist at Industrial and Commercial Bank of China International, believes that the ability of the Chinese state to exercise the coordination of key economic resources - chief amongst them financial capital - will prove to be its biggest comparative advantage when it comes to AI development.
He expects high-level coordination by the central government, in tandem with fast evolving energy and human resources endowments, to give China sufficient impetus to pull ahead of the US when it comes to leadership of the global AI sector.
Institutional advantages are the key to technological dominance
In a recent opinion piece entitled “China’s Comparative Advantages in the Era of Artificial Intelligence” (“人工智能时代的中国比较优势”), Cheng points to the crucial role that institutional features play in abetting the widespread adoption of new technologies.
“During the steam era, Britain’s advantages were derived from capital markets and its system of property rights,” he writes.
“In the era of electricity, the USA’s advantages came from its scientific research capabilities and corporate system.”
He expects the era of artificial intelligence to be no different, requiring its own set of unique institutional features to drive the profitable adoption of new innovations.
This will mean the development of a “highly organised innovation system,” whose chief features will include:
Cross-sector resource coordination,
The large-scale deployment of national computing infrastructure,
The coordinated development of overlapping industrial chains, as well as
Long-term financial support.
Cheng’s contention is that China’s unique brand of state capitalism is best equipped to rapidly assemble this suite of necessary institutional features.
“It’s China that possesses these system capabilities,” he writes.
“In a brief period of time, China is capable of deploying national computing networks, establishing clusters of large-scale data centres, driving reforms of the market for data as a factor of production, and forging cross-sector chain coordination systems.”
Control of financial capital is critical
Cheng believes the control that the Chinese state wields over the financial sector - giving it the ability to channel credit to sectors of priority importance - to be one of its most important tools in the competition for AI preeminence.
“The quality and effectiveness of financing for AI development will to a great extent determine whether AI technologies can transition from ideal to reality,” he writes.
In Cheng’s estimation, the foresight and prudence of China’s policymakers can overcome the shortcomings of financial markets when it comes to long duration investments.
“What’s most important is that China’s financial system can provide the ‘patient capital’ that AI requires,” he writes.
“AI is a classic case of an investment-intensive, long time-cycle sector with strong externalities…it’s hard to come by enough capital for investment by relying on the market alone.
“[China’s] policy banks, industry funds, government guidance funds and the multi-tier capital market jointly comprise a financial system that can adapt to such needs.”
Cheng’s remarks arrive just as Beijing steps up financial policies to channel more funds to the tech sector, in order to achieve scientific and technological “sovereignty” and wean Chinese industry off dependence upon the R&D of other nations.
Beijing is pushing for capital market reforms that will see regulators provide more overt support to the Chinese stock market, with a view to improving the equity financing prospects of promising tech firms.
This will include driving more “medium and long-term funds” to invest in stocks, as well as the use of China’s sovereign investors to play the role of “stabilisation funds,” stymieing excess price volatility with share purchases made at the behest of regulators.
Beijing is also using window guidance to drive commercial banks to lend to tech companies, as part of China’s “Five Great Chapters” of financial policy - the first and foremost of which is financing for scientific and technological innovation.
China’s new factor endowment advantages
Critics of Maoist era economic policies point out that China’s development strategy during the period was an extremely poor fit for its factor endowments.
From 1949 to 1978, China sought to follow the lead of the Soviet Union, focusing primarily on the development of heavy industry, while giving far less attention to the agricultural sector or the production of consumer goods.
This development strategy was highly capital intensive, yet did not require commensurate amounts of labour.
China’s factor endowments during the period were the opposite of those required by such a strategy, being characterised by a scarcity of capital, in tandem with a vast abundance of low cost labour.
Much has changed since the founding of the People’s Republic, however. Cheng opines that China’s current resource endowments give it another comparative advantage when it comes to the development of the AI sector specifically.
Clean energy
Cheng firstly highlights the critical role that abundant energy resources will play in the race for AI leadership.
“Whoever possesses a lower cost, more stable, and more scalable power system will have the stronger, underlying competitive advantage in AI,” he writes.
The high energy intensity of the AI sector is set to drive rapid growth in both data centre project and related power consumption moving ahead.
Cheng believes that China’s focus on the development of clean energy should put it in good stead when it comes powering the data centres that are indispensable to AI development.
“China holds a competitive advantage in the energy race for AI,” he writes.
“It possesses the world’s largest incremental capacity for renewable energy, an industrial chain system for converting energy into computing power, and the organizational capability to coordinate across regions and industries through state-level projects.”
China accounted for 61.2% (276.8 GW) of new global solar capacity and 69.4% (79.4 GW) of new global wind power capacity in 2024, according to figures from the International Renewable Energy Agency (IRENA).
This has translated into significant cost advantages, with China’s levelized cost of electricity (LCOE) for onshore wind power ($0.029/kWh) and for solar power ($0.033/kWh) both lower than the global average.
Beijing has also invested heavily in the transmission infrastructure needed to fully capitalise upon the country’s newly developed energy resources, with the development of UHV lines that connect climate rich regions in the west of China with hubs of demand on the eastern coast.
Cheng stresses the role of high-level coordination by the state in the development of these clean energy capabilities.
“This capability was not formed overnight, but rather is a structural advantage accumulated over the long-term, from power grid construction to energy structure adjustment, and from industrial policies to the project system.”
Human capital
Cheng points to China’s abundance of well-trained engineering talent as the other key resource that will power the rapid development of its AI sector.
He once again highlights the role of the state in the coordinated development of this critical factor endowment.
Top policymakers have driven the development of an engineering cohort with a focus on commercial applications, and the ability to fully capitalise upon the economic value of emerging technological innovations.
“Chinese engineers are highly integrated with the industrial chain, and possess a strong orientation towards applications as well as strong engineering capability,” Cheng writes.
“This enables technologies to be transformed into products more rapidly, incorporated into industries more rapidly, and validated and iterated into large-scale applications more rapidly.”
The scale of this talent pool is also considerable.
Figures from Georgetown University’s Center for Security and Emerging Technology indicates that by 2020 China had the world’s largest supply of engineers. It ranked first in terms of STEM graduates, exceeding the combined volume of the second and third-ranked nations.