iFlytek utilizes large AI models in pig farming

At the Changling Smart Farming Base at COFCO Joycome in Jilin province, a subsidiary of Chinese tech company iFlytek, iFLYHG Technology, is deploying artificial intelligence large models to transform traditional hog farming into a data-driven, intelligent operation.

China remains the world's largest hog producer and consumer. In 2025, the national hog slaughter volume reached 719.73 million heads, according to data from National Bureau of Statistics.

However, rapid expansion has brought rising management complexities, from disease prevention and feeding precision to environment control.

iFLYHG Technology said the new AI-powered system shifts the paradigm from experience-based to data-based management. Frontline expertise from veteran veterinarians and breeders is converted into algorithms. Vague empirical judgments become precise data-driven decisions, and manual barn patrols are replaced by intelligent unmanned monitoring.

Liu Pengchao, consulting director of smart farming at iFLYHG, said a collaborative model has been established: partners build computing and network infrastructure; iFlytek provides core AI capabilities such as multi-modal perception; and iFLYHG handles scenario implementation and system integration.

Key innovations include acoustic fingerprint models that detect abnormal sounds among barn noise, advancing disease early warning by two to three days; intelligent rail inspection robots for counting, weight estimation and temperature checks; and environmental control systems that adjust ventilation and temperature in real time. A precision feeding model tailors feed dosage to each hog's growth status.

The results have been notable. The PSY — piglets survived per sow per year — now steadily exceeds 29, ranking among the industry's top tier. In the farrowing house, one worker can manage nearly 800 piglets, significantly improving labor efficiency.

According to the company, the system has broken down data barriers among different equipment brands, enabling unified access and centralized analysis.

The practice at Changling demonstrates that a low-cost, replicable and scalable path for agricultural AI adoption is emerging — one that could inject new vitality into traditional industries across China, experts said.