A quiet yet profound transformation is unfolding in the fields on Beijing's outskirts: tractors operate without drivers, robots act as matchmakers for plant breeding, and robotic dogs diagnose crop health... These scenes, once reminiscent of the "cyber" aesthetic in sci-fi films, are becoming reality. Since the start of spring, visits to agricultural production sites in suburban Beijing have revealed a series of new spring farming scenes. From platforms to fields, and from laboratories to greenhouses, the landscape of agricultural technology is being redefined, with Beijing using technological innovation to reshape the future of farming.
Robots like "Ji'er" are at work. At the Fuxi Farm, an unmanned intelligent tractor operates autonomously. Inside the farm's command center, a large screen displays comprehensive information via satellite imagery. At the Shennong AI Farm, a robotic inspection dog patrols vegetable fields. Located in Huairou, this farm exemplifies advanced digital agriculture. Meanwhile, the Zhongguancun Agricultural Science Park Innovation Workshop in Pinggu serves as a hub for pioneering research.
**Innovation Hub: Pinggu Zhongguancun Agricultural Science Park** The slogan "Building the Agricultural Zhongguancun, Forging the Agricultural 'China Core'" stands out prominently at the innovation workshop in Pinggu's Yukou Town. Exhibition panels detail technological breakthroughs achieved here: the domestically developed "Fengxin Yi Hao" DNA chip for laying hens breaks foreign monopolies; the "Wode" series of white-feather broiler chickens fills a domestic void; and gene-edited pigs achieve 100% resistance to PRRS (Porcine Reproductive and Respiratory Syndrome). One by one, scientific challenges are being overcome.
Why has Pinggu become a "source" of agricultural technological innovation? The answer lies in strategic planning over recent years. In 2020, Beijing first proposed building an Agricultural Zhongguancun to establish an agricultural technology reform and innovation demonstration zone, complementing the city's goal of becoming an international science and technology innovation center. In October 2021, the Beijing Municipal Government and the Ministry of Agriculture and Rural Affairs signed a cooperation framework agreement to jointly build the China-Pinggu Agricultural Zhongguancun, elevating this strategy to a national level for agricultural technology development.
Pinggu's ability to shoulder this responsibility is backed by Beijing's unique and extensive research foundation. In the agricultural technology sector alone, Beijing boasts 24 agricultural research institutes and over 200 key laboratories and engineering technology research centers.
The core engine of the Agricultural Zhongguancun is the Beijing Jingwa Agricultural Technology Innovation Center, established in 2021. Jointly initiated by five entities including China Agricultural University, the Beijing Academy of Agriculture and Forestry Sciences, and the Beijing Shounong Food Group, the center breaks down long-standing barriers in agricultural technology, providing a platform to connect shelved research papers and patents with practical technical challenges in the fields.
Currently, the Jingwa Center's headquarters functional laboratories, along with demonstration parks for fruit, horticulture, and dairy, are largely completed. It has attracted expert teams from institutions like China Agricultural University and the Beijing Academy of Agriculture and Forestry Sciences, as well as R&D teams from agricultural technology innovation companies such as Sutuoke Technology and Aike Detection. The center has successfully introduced and incubated eight innovative small and medium-sized enterprises.
**Outlook: Agricultural Zhongguancun Enters a New Phase** Last October, Pinggu released the "Agricultural Zhongguancun Core Area Development 2.0" plan at the 2025 World Agricultural Science and Technology Innovation Conference. It clearly identifies synthetic biology and smart agriculture as core breakthrough directions, focusing efforts on building a national agricultural science and technology innovation center. This marks the entry of the Agricultural Zhongguancun into a new stage of development.
Currently, the main structure of the No. 4 Center project within the Agricultural Zhongguancun Comprehensive Research Center has been topped out. It is expected to be operational by 2027. Together with Centers No. 1 and No. 2, it will provide nearly 110,000 square meters of R&D space.
**Smart Farming: Fuxi Farm in Yujiawu, Tongzhou** At the Fuxi Farm in Tongzhou's Yujiawu area, 1,287 mu of experimental fields are uniformly planted with green wheat. Upon entering the field ridges, several large, silver-gray machines with a technological aesthetic come into view. "This is called the 'Honghu T200', an unmanned intelligent tractor," explains farm manager Wang Mengxuan. Its rear is equipped with a hitch to interchange agricultural implements according to the farming season. During the busy spring management period, the farm is nearly devoid of people. "On our farm, just a few people sitting in the office can manage these thousand-plus mu of land," Wang says with pride.
Inside the command center, a large screen displays the entire farm's information clearly on a satellite map. With a click of the mouse, the digital brain—the "Fuxi System"—appears. Clicking on any field instantly pops up detailed information such as soil moisture, crop growth status, and pest warnings.
"In the fields, there are many intelligent facilities: plant pathogen spore traps, soil moisture sensors, IoT pest monitoring devices, etc. They act as sensitive 'noses' and 'eyes', providing 24/7 monitoring data for the digital brain," Wang explains. The system not only enables field perception, data analysis, and precise execution but, more importantly, it can also make decisions.
Should winter wheat receive additional fertilizer? Is pest control needed? Previously, this relied entirely on experienced farmers. Now, AI can generate professional spring management plans within seconds. When operations are required, staff simply issue a command with one click, and the corresponding intelligent agricultural machinery automatically starts and plans its route to work in the fields.
**Background: Modern Farm Built in Just Half a Year** The Fuxi Farm is a pilot project by the Chinese Academy of Sciences for developing smart agriculture. It aims to translate cutting-edge technologies like big data and artificial intelligence into practical agricultural productivity, enhancing production efficiency while providing strong support for food security. In early 2025, the Beijing Municipal Bureau of Agriculture and Rural Affairs began introducing the Chinese Academy of Sciences' "Fuxi Farm" project. After multiple surveys, the location was finalized in Yujiawu Township, Tongzhou District. The Beijing "Fuxi Farm" was constructed and opened to the public in less than half a year.
**Intelligent Breeding: Shounong Cuihu Innovation Workshop in Shangzhuang, Haidian** In the intelligent tomato production and breeding greenhouse at the Shounong Cuihu Innovation Workshop in Haidian's Shangzhuang Town, a white robot by the racks immediately catches the eye. It is square-shaped with a freely extending and retracting mechanical arm, slender like an enlarged embroidery needle. "This is 'Ji'er'," says Xu Cao, a researcher at the Institute of Genetics and Developmental Biology, Chinese Academy of Sciences.
"Ji'er" (GEAIR) is the world's first intelligent breeding robot capable of automatic navigation and precise hybrid pollination, jointly developed by research teams from the Institute of Genetics and Developmental Biology and the Institute of Automation, Chinese Academy of Sciences.
Why use a robot as a "matchmaker"? This relates to hybrid breeding. Traditional hybrid breeding first requires "emasculation"—removing the stamens from the maternal flower. Then, using tweezers, pollen from the paternal plant is applied to the maternal stigma. The entire process is intricate and delicate, more meticulous than embroidery, and entirely reliant on manual labor. Some crops with great potential (like soybeans) have been unable to leverage the advantages of hybrid breeding on a large scale due to pollination difficulties and high labor costs.
Xu Cao's team tackled the problem in two steps. First, modifying flower structure: using gene-editing technology to make stamens naturally dehisce and pollen become sterile during early growth, causing the stigma to naturally protrude, successfully creating a "robot-friendly" structural male sterile line. Second, building the robot: "Ji'er" relies on deep learning algorithms. After training its neural network with 12,800 images, it can accurately identify flowers and determine stigma orientation, then use its mechanical arm to complete pollination.
"Ji'er's" operational efficiency is astonishing. Through breeding practice and improvements, pollination of a single flower now takes as little as 10 seconds, with stigma recognition accuracy reaching 90% and a single cruise pollination success rate exceeding 82%. It can cruise and pollinate repeatedly around the clock, ensuring successful fruit set. In soybean breeding, it has for the first time enabled the rapid creation of structural male sterile soybean lines, saving 76.2% of manual pollination time. It holds promise for breaking through hybrid breeding challenges in soybeans, increasing yield per unit area, and alleviating China's long-term dependence on soybean imports.
**Outlook: Developing Embodied Intelligent Agricultural Robots** Currently, Xu Cao's team is focusing research on developing embodied intelligent agricultural robots that integrate functions like agricultural biological breeding, field management, monitoring, and harvesting/transportation. This aims to solve application challenges for intelligent robots in the complex and variable agricultural environment. The goal is to have embodied intelligent robots handle laborious tasks people don't want to do, delicate tasks people find difficult, and dangerous tasks people shouldn't do, taking over the entire process from seed to seed.
**Digital Plant Protection: Shennong AI Farm in Huairou** On the field ridges, a nimble "robotic dog" moves on four legs, steadily navigating through vegetable plots. It walks and stops intermittently, "looking" at tomatoes one moment and "observing" lettuce the next. While seemingly aimless, it is actually conducting serious "field inspections." On farmers' mobile screens at the field edge, data on crop growth and whether they are "sick or infested" updates in real-time.
At the "Shennong AI Farm · Huairou Yanqi Base", an intelligent inspection robotic dog has become the most capable "new farmer". During daily inspections at the base, it can quickly identify crop types and maturity, with a pest and disease recognition accuracy rate exceeding 90% and a real-time response time of less than 0.2 seconds.
Wang Yaojun, head of the China Agricultural University Shennong Large Model team, explains that the robotic dog is equipped with the high-intelligence recognition system "Shennong Pest and Disease Identification Agent", integrating AI vision and intelligent decision-making technologies. It can not only "see" the affected areas but, like an experienced agricultural technician, also "understand" and "diagnose" the type and severity of pests and diseases. Besides the robotic dog, there are also small pest identification and diagnosis devices for precisely recognizing tiny pests like thrips, and pest sex lure intelligent identification devices for pest monitoring and trapping. They respectively undertake the three major tasks of "inspection and diagnosis", "precise identification", and "trapping and monitoring".
After multiple rounds of upgrades, the system can now cover 75 crops including wheat, corn, and peaches, intelligently identifying 615 types of pests and diseases. It contains over 50,000 pieces of pest control technical information, 400,000 high-quality labeled plant protection image data points, and 30,000 high-quality plant protection Q&A data points. This breakthrough makes it the first AI plant protection agent in China's plant protection field to systematically integrate identification, early warning, and control functions.
**Promotion: Shennong Large Model Opens to Global Users** This powerful "field doctor" is bringing Beijing's intelligent plant protection directly to field ridges and greenhouses. Wang Yaojun states that the team has initiated the construction of a research center focused on artificial intelligence agents for tropical fruits and crops at the China Agricultural University Sanya Research Institute in Hainan. They plan to expand the system's coverage by 50 more pest and disease types and use this as a pivot to serve overseas pilot bases in Southeast Asia. During this spring's farming season, the system has already provided service support to farms in Heilongjiang and is currently conducting real-time monitoring of wheat growth in Henan. Just this month, the team also began cooperation with the UN Food and Agriculture Organization (FAO) to provide plant protection services for smallholder farmers in Kenya, with pilot projects expected in some areas of the country within the year. Adhering to the concept of openness and sharing, the Shennong Large Model is open to global users, supports 10 languages, and currently has users across 15 countries and regions.
Comments