China’s rapid adoption of artificial intelligence is not yet translating into stronger profitability for listed companies, highlighting the difficulty of turning technological investment into sustainable economic growth.
Yan Shen, a professor of economics at Peking University‘s National School of Development, said research presented at the 2026 Per Jacobsson Lecture found that greater AI use was associated with higher productivity on average, but not with measurable improvements in return on assets or return on equity among the Chinese listed firms studied.
Prof. Shen described the results as a paradox. AI adoption is expanding rapidly and creating new commercial opportunities, yet the financial returns remain uncertain.
She said the findings also showed that companies with higher productivity tended to benefit more from AI, potentially widening the gap between stronger businesses and those struggling to adopt the technology effectively.
China had about 700 million generative AI users in the first half of 2026, according to figures Prof. Shen attributed to the China Internet Network Information Center. She also cited an adoption rate of about 30% among large enterprises meeting the relevant threshold in the data she presented.
Computing demand was rising rapidly, with Prof. Shen reporting year-on-year growth of about 170% in computing power and a roughly fourfold increase in demand during the first quarter of 2026. She also cited a substantial increase in daily token consumption compared with March 2025.
Despite that expansion, Prof. Shen identified three reasons businesses may struggle to earn returns from AI investment.
Companies may not know which AI tools are best suited to their operations. Some adopt the technology because competitors are doing so, rather than because it has a clear business case. Others risk seeing their investments become obsolete before they can generate sufficient returns, as AI capabilities evolve quickly.
The result is a potential mismatch between rising technology spending and the income businesses can generate from it.
Shen also warned that AI could intensify China’s deflationary pressures by allowing companies to produce goods and services at lower cost without a corresponding increase in household incomes and demand.
The effects on workers are also mixed. Research she presented found that households in occupations more exposed to AI experienced lower labour income for the household head and a higher probability of carrying debt.
However, Prof. Shen pointed to opportunities in healthcare and consumer markets, where AI can improve service coverage, reduce costs and help businesses identify unmet demand.
Her findings suggest that widespread AI use alone may not guarantee stronger economic growth. The gains will depend on whether businesses can convert productivity improvements into profits, workers’ incomes and sufficient consumer demand to support further expansion.
