On February 4th, during the ongoing Sichuan Provincial Two Sessions, the 2026 Provincial Government Work Report mentioned the need to accelerate the cultivation of emerging and future industries, and to strategically plan for industries such as quantum technology and embodied AI. Currently, Embodied AI has advanced from the laboratory to the forefront of industry. According to the "Embodied AI Development Report (2025)" jointly released by the China Academy of Information and Communications Technology and the Department of Electronic Engineering at Tsinghua University, by December 2025, China's Embodied AI and robotics sector had witnessed 744 investment events, with a total financing amount reaching 73.543 billion yuan. So, what is Sichuan's position in this cutting-edge field that merges the physical and the intelligent? How can it leverage its strengths and circumvent its weaknesses to forge a distinctive development path? A reporter interviewed two provincial committee members specializing in technology, who offered forward-looking insights into Sichuan's Embodied AI development path from the dual perspectives of core technological insight and industrial ecosystem construction.
Provincial Committee Member Li Jianqing, a professor at the University of Electronic Science and Technology of China's School of Electronic Science and Engineering, and Provincial Committee Member Zhang Jianhua, Dean of the School of Computer Science and Artificial Intelligence at Southwest Minzu University, provided their analyses. Professor Li Jianqing's perspective forms the first part of the analysis.
Part 1: The "Sichuan Strategy" under Technological Stratification: Entering Vertical Battlefields with a Complete Chain. As a provincial committee member from the front lines of academic research, Professor Li Jianqing from the University of Electronic Science and Technology of China's School of Electronic Science and Engineering has long focused on emerging industries in the technology sector. In his view, the development of embodied intelligence can be broadly divided into four layers: the application layer, the algorithm layer, the hardware layer, and the chip layer. Examining the national landscape through this framework reveals a clear competitive situation for Sichuan. "At the chip layer, we might not match Shanghai; at the hardware layer, we face direct competition from enterprises in the Yangtze River Delta, such as Zhejiang's Unitree Robotics; at the algorithm layer, particularly for large models, Beijing holds significant advantages." However, he believes Sichuan's true strength lies in possessing a relatively complete industrial chain spanning chips, hardware, algorithms, and application scenarios. Therefore, the optimal path is not to engage in head-on competition with stronger rivals at any single technological layer, but rather to focus on one vertical domain—such as cultural tourism, healthcare, or emergency response—and commit to "saturated" investment. By leveraging the local industrial closed loop, Sichuan can build comprehensive chain competitiveness, from scenario definition and algorithm optimization to hardware adaptation.
Professor Li further illustrated this viewpoint with an industry example: "For instance, Unitree Robotics' robots can perform highly difficult actions like backflips and kicking watermelons, which precisely reflects the capability of integrated full-chain development." However, he noted that China's embodied AI industry still suffers from a "fragmented chain" phenomenon: algorithm, hardware, and chip R&D are dispersed, and core computing chips largely rely on imports. "This model leads to lower innovation efficiency and constrains the deepening of competitiveness." So, which "vertical battlefields" must Sichuan prioritize? Professor Li outlined three highly promising directions: In cultural tourism, "robot ambassadors" capable of guiding and deep interaction could emerge; in healthcare and wellness, there is an urgent demand for home care robots; particularly crucial is disaster prevention, mitigation, and emergency response—"Sichuan's complex terrain and high disaster risks make deploying robots for inspection, monitoring, and initial handling in high-risk environments, replacing human personnel in dangerous situations, of immeasurable social value." When asked what Sichuan should prioritize for breakthrough, his answer was chip development. "Chips are the foundational bedrock of all information industries, and Sichuan already has a certain base in this area."
Part 2: Developing the "Sichuan Path": From a Solid Foundation to Distinctive Breakthroughs. Complementing Professor Li's in-depth technical analysis, Provincial Committee Member Zhang Jianhua, Dean of the School of Computer Science and Artificial Intelligence at Southwest Minzu University, stated, "Sichuan has a solid foundation across the entire chain for developing embodied AI." He listed four major advantages: First, synergistic industrial pillars: The three dominant industries of electronic information, equipment manufacturing, and artificial intelligence can provide integrated support encompassing "computing power + hardware + algorithms + scenarios." Second, abundant scientific and educational resources: Institutions like Sichuan University and the University of Electronic Science and Technology of China have profound expertise in robotics and brain-computer interfaces, while platforms like the Tianfu Laboratory and the Chengdu Humanoid Robot Innovation Center are accelerating technology transfer. Third, unique scenario provision: Chengdu benefits from the overlapping policies of being a National AI Innovation Pilot Zone and an Application Leading Zone, opening up testing scenarios in fields like healthcare, industry, cultural tourism, and the low-altitude economy. Fourth, strong government support: Provincial-level special projects, Chengdu's ten-billion-yuan industrial fund, along with policies like "computing vouchers" and "subsidies for the first set of equipment," effectively reduce R&D and commercialization costs for enterprises.
The fertile ground is already sprouting new shoots. Dean Zhang shared an innovation case from a local enterprise: Chengdu Ruixinxing's development of the world's first training-free industrial embodied picking robot, which has already been deployed in 12 manufacturing companies, solving the industry pain point of traditional robots requiring retraining for every production change. Based on his deep observation of the industry, he suggested that Sichuan's exploration of application scenarios should be layered and focused: Priority should be given to promoting the "digital transformation and intelligent upgrading" of industrial manufacturing and modern agriculture, such as training-free picking, flexible assembly, and precision inspection. Next, efforts should concentrate on the big health sector. Furthermore, it should empower the integration of culture and tourism, developing intelligent guided tours, security patrols, etc. Finally, opportunities can be explored in ethnic regions and specialized operations, such as low-altitude logistics, power operation and maintenance, ecological inspection, and bilingual services.
Also not shying away from current challenges, Dean Zhang pointed out that the industry currently lacks pilot-scale testing platforms suited to various sectors, making the transition from laboratory technology to mass production difficult; a disconnect exists between university R&D and enterprise needs; reliance on imported core components (such as servo motors) leads to high costs; and there is a scarcity of interdisciplinary talent proficient in both algorithms and engineering. To address these, he outlined a "Sichuan Path" with distinct characteristics: Sichuan should avoid blindly chasing the red ocean of humanoid robot整机 (whole units), and instead delve deeply into niche sectors where it possesses local advantages, such as industrial flexible manufacturing, modern agriculture, supporting industries for the low-altitude economy, and healthcare/wellness. On this basis, it should focus on shaping technological differentiation, forging cutting-edge directions like "training-free operation, emotional interaction, brain-computer interfaces" into unique core strengths that others lack. Furthermore, mechanism innovation is needed, implementing an open model of "scenario posing problems - enterprises solving them - government providing subsidies," proactively releasing unique scenario resources in fields like aviation and healthcare. Simultaneously, it should actively build an integrated ecosystem of "universities + research institutions + leading chain enterprises + pilot-scale platforms + scenario providers,"打通 (opening up) the critical pathway for results transformation. Ultimately, by deepening industrial coordination and resource sharing within the Chengdu-Chongqing economic circle, the region can jointly create a hub for embodied AI innovation and a highland for the industry in Western China.
Comments