AI News Feed
Market watch
Robotics

Robots Trade Stage Tricks for Real Jobs at Yibin and Shanghai Events

At Yibin's robot job fair and Shanghai's Outer Bund, robots must pass real work tests, showing AI's shift from stage stunts to deployment.

At the Yibin event, the “Jiangyuan Cup” embodied-intelligence scenario competition, held at the Yibin International Convention Center during the World Power Battery Conference, drew more than 130 registered teams across industrial, commercial, lifestyle and open-innovation tracks. According to a QbitAI report, the contest used no staged routines: robots had to move brewery-style grain bags, sort and assemble parts with positioning tolerance of less than 0.1 millimeter, patrol industrial areas with multi-sensor fusion and carry out delivery runs across flat ground, gravel, slopes and stairs.

Organizers framed the exercise as a recruiting and verification loop. Teams brought physical machines rather than proposals, and industrial partners asked about pricing, delivery times and customization. Qingzhi Capital has already pushed two projects to pilot in Yibin, including a tea-making robot working with Sichuan Tea Group and an embodied-hardware design platform linked to a local university. Yibin's Southwest Embodied Intelligence Training Center, opened in September 2025, plans to operate more than 250 data-collection robots and produce over 200,000 hours of multimodal interaction data this year.

At the Outer Bund Conference, Lingbo Robotics displayed a working “smart pharmacy” developed with Shanghai Guoda Pharmacy. According to a Leiphone report, a robot receives orders, moves to shelves, picks a medicine package and delivers it to a pickup counter, assisting pharmacists with night-shift sorting. Real stores have 80-centimeter aisles, transparent packaging, reflective surfaces and randomly placed goods, but the system requires no store refitting.

The pharmacy robot and other logistics machines at the booth run on Lingbo's LingBot-VLA 2.0, described as an “embodied-native general-purpose brain.” During pre-training, the model has adapted to 17 robot brands and 20 hardware configurations, covering single-arm, dual-arm, bipedal and wheeled robots. At the booth, the same brain controlled different robots in logistics and industrial tasks, an attempt to escape the old pattern in which changing a body or gripper required rewriting control code.

Lingbo's chief scientist, Shen Yujun, was scheduled to join a main-forum roundtable on “physical intelligence” and discuss the industry's shift from building bodies to building brains. The company also said it would launch an embodied large-model challenge along with ModelScope and Alibaba Cloud's Tianchi platform, opening its technology and scenarios to outside developers.

Also at the Outer Bund Conference, Taichu Yuanqi displayed its domestic AI-for-Science computing platform and publicly unveiled TecoWeatherNext, a typhoon tracking system. According to a QbitAI report based on company materials, the platform went online in July and is built on Taichu's self-developed heterogeneous many-core AI chips. It supports Loongson, Sunway, Phytium and x86 CPUs, and covers the whole stack from hardware to acceleration libraries. Taichu algorithm engineer Du Feifan was quoted as saying that the typhoon system fuses meteorological data, AI forecast models and tracking technology to show storm trajectories and support disaster prevention.

Taichu used the conference's tech job fair to open more than ten R&D positions in AI engineering, chip design and basic software. Campus recruiter Liu Ruihan said the shortage of domestic computing talent has grown as China's AI sector expands, and the company wants to bring more engineers into domestic chip and computing product development. Taichu said more than 70 percent of its staff hold master's or doctoral degrees, and it has hired over 50 campus recruits in Beijing, Shanghai and Wuxi in the past two years.

In Yibin, the competition's trophy was described as becoming an order for the winning teams. In Shanghai, Lingbo's pharmacy robot is already serving customers at Guoda stores. In both places, the decisive question is whether a robot can hold a shift under real conditions, not how well it performs on a stage.