ACE ROBOTICS today unveiled Kairos 3.1, its latest action-oriented world model, alongside Ambient Capture Engine 2.0 and three commercial solutions spanning instant retail, hospitality and outdoor service scenarios. The announcements were made at a forum during the 2026 World Artificial Intelligence Conference, focused on advancing physical AI from "understanding" to "execution." The event brought together economists, technology executives and embodied AI researchers to examine how general-purpose world models can move from laboratory research to industrial deployment.
The forum also marked the launch of PHYSICAL IQ, a unified benchmark for embodied physical intelligence, jointly initiated by the Shanghai Artificial Intelligence Association, the Shenzhen Loop Area Institute and the East China branch of the China Academy of Information and Communications Technology, with participation from more than 20 universities and industry partners. Nobel laureate Thomas J. Sargent discussed the limitations of current intelligent systems in rare and previously unseen situations.
Kairos 3.1 is designed as a natively unified model that integrates generative, physical and cognitive intelligence within a single architecture, rather than combining separately developed capabilities. Built on a hybrid Transformer architecture with a shared mixed-attention mechanism, it brings visual observations, language instructions, force and tactile signals, and policy trajectories into a unified latent space. This supports an "understand, reason, execute and reflect" loop, enabling robots to break down long-horizon tasks, simulate physical cause and effect, and adjust strategies in real time.
At the core of its spatial understanding is ACE-BRAIN-0.5, which achieved state-of-the-art results across 12 public evaluations covering spatial understanding, navigation, manipulation and task-progress assessment. For physical generation and reasoning, Kairos-HomeWorld supports whole-home scene generation and object interaction, built on 300,000 residential floor plans and 5,000 simulated home environments. In internal testing, the Kairos 3.1 8B model achieved an inference latency of 125 milliseconds on the NVIDIA Jetson Thor platform.
ACE ROBOTICS also introduced the "Information-Density Law" for embodied models, classifying data across five levels from L1 to L5. Higher-density data incorporates force-tactile signals and failure-recovery trajectories, supporting stronger generalization and continuous learning. Ambient Capture Engine 2.0, built on this principle, includes ACE Ego Kit with the ACE Sense Glove offering 0.01 newton sensitivity, ACE Data Engine for automated annotation, and ACE Ego Matrix to standardize data across systems. The company also open-sourced ACE-Data-0, an L5 household-interaction dataset.
Three commercial solutions were introduced: Xiaoman for instant retail with the W1 fulfillment robot, Xiaoxin for hotel laundry operations, and Xiaotu for outdoor autonomous service using a "one brain, many bodies" architecture. Xiaoman has been deployed with customers including Sense MartGo, Kuaikeda and PetroChina convenience stores, with plans to expand to 10,000 stores within two years. Xiaotu has been deployed at cultural-tourism events like the Begonia Flower Festival in Tianjin.
ACE ROBOTICS also announced partnerships with the Caohejing Development Zone, PetroChina, Kuaikeda and SenseTime Shanhui, and launched the World Model Cloud Ecosystem initiative with Baidu AI Cloud, Alibaba Cloud, Huawei Cloud, Tencent Cloud and SenseCore AI Cloud. By connecting high-density data, unified world models, robot platforms, commercial applications and shared evaluation standards, ACE ROBOTICS aims to support the transition of embodied AI from demonstrations to sustained real-world operations.
