China’s industrial robot boom is spilling beyond its own factory floors, raising a question for developing economies: will the machines arrive as another import, or help build industrial capabilities of their own?
Industrial enterprises above a designated size produced 537,689 robots in the first half of 2026, a 28 per cent increase from a year earlier. China also exported Rmb6.29bn of industrial robots, up 18.6 per cent, to 141 countries and regions, a customs official said in July.
The figures measure output and export value, rather than the number of machines shipped. But they illustrate the scale behind China’s push overseas. The country already accounted for 295,000 installations, or 54 per cent of the global total, in 2024. UN trade records list India, Vietnam, Thailand, Mexico and Brazil among destinations for Chinese industrial robots that year.
Trade data cannot show what happened after the machines arrived. Two factory projects offer a more useful, if incomplete, picture.
At AutoAlliance Thailand, a Ford-Mazda joint venture in Rayong, JAKA worked with local integrator AppliCAD to deploy collaborative robots for handling, assembly, screwdriving, gluing and inspection. Supplier-linked coverage put the deployment at more than 50 cobots in 2025. JAKA said in April that it had grown to more than 85.
The expansion suggests that the project moved beyond a limited trial. But the supplier has not disclosed independent figures for productivity, defects, safety or employment.
In Pune, Jendamark India integrated a Chinese DOBOT cobot, a gauging machine and a SCARA robot into a differential-case assembly line at Kalyani Technoforge. A case study by the integrator attributes a 40 per cent reduction in cycle time and a 70 per cent improvement in quality to the project. It says automation reduced operator fatigue and allowed workers to move towards quality control and other higher-value tasks.
Those results apply to the three-part system, not to the Chinese cobot alone, and have not been independently verified. Jendamark does not give the commissioning date, uptime or change in headcount.
The projects point to where much of the value in automation is created — the same question that decides whether a cross-border deal transfers capacity or only sells output. AppliCAD occupied the integration layer in Thailand. Jendamark built interfaces, staged the Indian installation and connected the equipment to its own production software. Such work determines whether a machine can fit an existing factory and continue running. Neither account, however, reveals how much technical access, expertise or recurring revenue remained with local companies.
A new initiative may bring that question to more markets. The Global Robot Application Exploration Program, launched on August 20 by the World Robot Cooperation Organization and the Chinese Institute of Electronics, plans to select 1,000 application scenarios a year for free robot trials and further development. Its published description says feedback from real-world use will guide product development.
The programme is not limited to industrial robots or developing economies. Yet its structure exposes an imbalance in the proposed exchange. Suppliers gain access to new industries, operating conditions and product data. The return for host economies is less clearly defined.
According to reporting by China Economic Net, industry representatives from developing countries argued that equipment should be paired with training and local deployment support. The report described proposals for local-company maintenance and software networks, workers able to integrate and program systems, and research linked to national deployment strategies.
The distribution of gains will matter. A World Bank study covering five Southeast Asian economies estimated that industrial robot adoption created about 2mn jobs for skilled formal workers between 2018 and 2022, while displacing 1.4mn lower-skilled formal workers in routine manual jobs. It did not isolate Chinese robots, but its findings make training and worker transition central to any industrial bargain.
The new programme’s operating terms were not public as of August 26. It has not said who will pay for transport, integration, certification, maintenance and downtime; whether local companies will receive documentation and diagnostic tools; who will control operating data and locally developed intellectual property; or what happens to equipment after a trial.
Free use may lower the cost of experimentation. It is not, by itself, technology transfer. The more consequential measure will be what remains when the trial is over: factories able to run the machines, local companies paid to adapt and maintain them, and workers equipped to move into more productive roles.
