At a digital government forum in Guiyang in August 2026, the director of China's Agricultural Information Center set out how he expects smart agriculture to develop over the coming phase.

He framed the field around three elements, sensing, decision-making and execution, supported by three foundations: computing power, models and data. Agricultural data, and biological data on crops and livestock in particular, remains genuinely scarce, he noted. A large model, in his view, only becomes true artificial intelligence when paired with an agent that acts on its output.

From there he described three shifts: from technical feasibility to proven value; from answering questions to completing tasks, which requires machines to physically fertilize, spray and harvest; and from isolated pilot projects to full-chain integration.

The practical implication is that the execution layer is where much of the next round of value will be created. As decision systems mature, the hardware that carries out those decisions, valves, pumps, dosing units and controllers, determines whether the intelligence ever reaches the crop.

Source: Wang Xiaobing, Director of the Agricultural Information Center, Ministry of Agriculture and Rural Affairs of China, speaking at the Digital Government session of the China International Big Data Industry Expo, reported 27 August 2026. Original report: https://www.toutiao.com/article/7679117022164238883/