Ant Lingbo Signs Longsheng Weirui, Dema Technology and Others to Bring Embodied AI Brain to Production Lines

Under the agreements, Ant Lingbo will act as a core embodied model provider in industrial scenarios and will participate in building the Longsheng Weirui Embodied Intelligent Robot Innovation Center. The first phase will focus on industrial loading and unloading, with future plans to provide intelligent solutions to global industrial manufacturing leaders including Bosch and Baosteel. Ant Lingbo will also work with De'er Xiangwang to explore the digital and intelligent transformation of automotive manufacturing workshops.

In logistics, Ant Lingbo will support Dema Technology and other companies through engineering enablement, secondary development collaboration and testing of new technologies. The partners will tackle high-speed flexible sorting of massive parcels and explore more diversified deployment scenarios to improve practical results.

The agreements add to Ant Lingbo’s previous work. The company has already run a continuous medicine-picking “smart pharmacy” solution in offline pharmacies. At the recent Global Digital Trade Expo, it reached cooperation with the Arab Federation for Digital Economy (AFDE) on technical models, localized scenario adaptation and industrial ecosystem integration, opening a new phase for large-scale deployment of its “one brain, multiple machines” approach in the physical world, according to the report.

At the ongoing 26th China International Industry Fair, Ant Lingbo is showing partners and applications including precise pharmacy picking, flexible sorting in smart logistics and loading and unloading on industrial production lines. The demonstration includes two robot bodies with different configurations from Taihu and Xingchen Intelligence, both equipped with the LingBot-VLA 2.0 brain, completing high-precision industrial loading and unloading and flexible sorting tasks.

Adaptation cost has long been a key threshold for embodied intelligence to enter industrial applications, according to the report. Moving a robot from one scenario to another often requires new data collection, model training and system debugging; replacing hardware can lead to further adaptation work. Insufficient model generalization and differences in robot hardware make it difficult to copy single-point validation directly to more production lines.

Ant Lingbo CEO Zhu Xing said the company has moved away from the traditional approach of tuning individual machines and instead trained an embodied foundation model for the physical world from scratch to create a native general-purpose “brain” for robots. Its self-developed LingBot-VLA 2.0 foundation model was already adapted to 20 robot configurations from 17 brands during pre-training, setting what the company calls a mature “one brain, multiple machines” engineering benchmark in the industry.

As of now, Ant Lingbo’s open-source models have received 32,900 stars on GitHub, ranking among the leading domestic embodied models, the report said. Overseas mainstream technical communities have also adapted Lingbo models. LingBot-VLA and LingBot-VA have completed ecosystem adaptation for LeRobot, described as the world’s largest robot development platform, while LingBot-Vision has been merged into the mainline of timm, a mainstream visual backbone model library. LingBot-World became the world’s first world model supported by both vLLM and SGLang, two mainstream open-source inference engines.

From narrow aisles in smart pharmacies to production line operations at the industry fair and dense orders in industrial manufacturing and smart logistics, Ant Lingbo is using a dual-drive strategy of leading models plus engineering tools to work with industry partners on practical problems in real scenarios and apply “one brain, multiple machines” to more real business cases, the report said.

Editor's Summary

Ant Lingbo, Ant Group’s embodied intelligence unit, has signed strategic partnerships with Longsheng Weirui, Dema Technology and De'er Xiangwang to apply its general-purpose embodied brain in industrial and logistics settings. The company is showcasing pharmacy picking, logistics sorting and industrial loading and unloading at the 26th China International Industry Fair while promoting its LingBot models and “one brain, multiple machines” approach. The deals and demonstrations mark a push to move embodied AI from single-scenario trials toward broader commercial deployment.