EEmbodied AI Hub
IndustryDeploymentEditor’s pick

Xingchi Power and PKU Healthcare: A Pragmatic Path for Embodied AI in Hospitals

Xingchi Power and PKU Healthcare are collaborating on a three-step pipeline—simulation training, scenario validation, and hospital deployment—to bring embodied AI into medical settings safely. The partnership leverages Xingchi's world model engine and data headset to create high-fidelity virtual hospitals, addressing the lack of training environments and expert data. This pragmatic approach prioritizes safety over speed, focusing on tasks like medication sorting and delivery.

simulation trainingmedical roboticsPKU HealthcareXingchi Powerfeed:leiphoneworld modeldeploymentleiphone

Source: 雷峰网 · July 28, 2026

Share this article so more people can see it

Embodied AI entering hospitals isn't about plopping a robot into a ward. It requires understanding the hospital, the staff, and a step-by-step progression from simulation to validation to deployment. That's exactly what Shenzhen-based Xingchi Power Intelligent Technology and PKU Healthcare Management are doing.

Xingchi Power, a full-stack embodied AI model company, and PKU Healthcare, which provides real clinical scenarios and medical expertise, have announced a deep collaboration. They aim to build a complete pipeline from "simulation training" to "scenario validation" to "hospital application." The core bottleneck they are tackling: no environment to train in, no expert teachers to learn from.

Their approach is refreshingly pragmatic. Instead of rushing a robot into a hospital, they are first building a "high-fidelity virtual hospital" using Xingchi's KAI World Model engine. This engine recreates typical medical settings like wards and operating rooms, allowing robots to iterate and fail safely before ever touching a real patient.

To capture expert techniques, Xingchi uses its KAI Halo first-person data collection headset. Real medical staff wear the headset during procedures, and their actions are recorded and processed into high-quality multimodal training data via Xingchi's Embodied AI Infra. This effectively gives robots a team of "gold-medal teachers."

The second step is training in the virtual hospital. Xingchi builds an interactive medical digital asset library from real images and videos of drugs, instruments, and storage equipment. This allows for low-cost, repeated testing of tasks like sorting and grasping without consuming physical supplies. Researchers can freely combine assets to recreate various task scenarios, enabling massive-scale operational learning. Robots accumulate "muscle memory" through countless virtual trials, correcting errors and building reliability.

Only after rigorous simulation training do they move to scenario validation. In environments that closely mimic real wards and operating rooms, the system is tested for stability, reliability, and safety. Only then is a small-scale pilot considered. The initial focus is on "capillary" tasks of smart hospitals: test tube sorting, medication delivery, and bed transport. These are low-risk, high-frequency tasks that are ideal for gradual validation.

For founders and operators, this partnership offers a template. The medical domain is unforgiving; safety is non-negotiable. Xingchi Power and PKU Healthcare are not promising overnight revolutions. They are building a foundation—data, simulation, validation—that can scale. The key takeaway: don't skip steps. Invest in simulation environments, capture real expert data, and validate in controlled settings before deployment.

The collaboration also signals a shift in how embodied AI companies approach verticals. Rather than a general-purpose robot, Xingchi is tailoring its stack to medical specifics: world models that understand hospital layouts, data pipelines that respect privacy and precision, and validation protocols that meet clinical standards. This vertical focus could be a competitive moat.

Looking ahead, the partners plan to deepen their work on scenario data, world models, and simulation training. The goal is to steadily embed embodied AI into medical practice, providing a reliable engine for smart hospitals and high-quality healthcare services.

Source: 雷峰网.

Discussion

Tell us what you think — comments make stories more useful for builders and founders.

Tell us what you think!

Xingchi Power and PKU Healthcare: A Pragmatic Path for Embodied AI in Hospitals

Have an account? Log in to use your display name and avatar.

Email is optional and never shown on the page.

More insights

IndustryDeploymentFeatured雷峰网

Video Generation Is Not Planning: Georgia Tech's Danfei Xu Proposes Factor Graphs to Bridge the Action Gap

Danfei Xu argues that high-fidelity video generation does not equal physical planning. He proposes compositional world models using factor graphs and temporal attention to enable test-time skill composition, improving success on out-of-distribution tasks. The approach reveals a video-action gap where action models lag behind visual generalization.

Read
IndustryDeployment雷峰网

RSS 2026 Debate: Data Source, Control Modality Split Robotics Community

At RSS 2026, five experts debated whether world models and policy models should be unified or separate, whether control should rely on pixels or actions, and whether training data should come from the internet or embodied robots. The 'separate' camp won the first round, 'pixel' vs 'action' ended in a draw, and 'internet data' won the third round 61% to 39%. Key takeaways: modular design is preferred for debugging, video alone is insufficient for contact-rich tasks, and hybrid data strategies are

Read
IndustryDeploymentFeatured雷峰网

RSS 2026 Test of Time Award: Morphology as Computation Resets Embodied AI Priorities

The RSS 2026 Test of Time Award for the RBO Hand validates that passive mechanical design can replace compute-heavy sensor suites, signaling a strategic pivot for embodied AI startups. The award highlights environmental constraint exploitation and extreme cost advantages, but high-speed dynamic deployment remains unproven. Founders should reconsider hardware-first approaches over pure algorithmic scaling.

Read