XTRON Robotics closes an A1 round to put embodied AI into hazardous plants

XTRON Robotics, a builder of embodied-intelligence systems for hazardous industrial settings, said on 1 September it had closed an A1 funding round worth tens of millions of yuan.

The round was backed by Qianhai Ark, Xihu Science and Innovation Investment and Guangyang Co. The proceeds will go to refining its XBOT-AEGIS embodied-intelligence system, developing multi-form explosive-proof robots, and expanding at home and abroad.

XTRON explosive-proof inspection robot operating at an industrial plant
XTRON’s explosive-proof inspection robot built for oil, gas and chemical sites. (Source: OFweek)

Founded in December 2022, XTRON focuses on special-risk process industries such as oil and gas extraction, chemical production and nuclear maintenance. Its founding team comes from Harbin Institute of Technology, with technical backing from HIT’s robotics institute and the Fujian (Quanzhou) advanced manufacturing research institute.

Founder and chief executive Fu Zhe spent a decade in industry and holds dual bachelor degrees from HIT and the University of Birmingham and dual master degrees from the University of Sheffield and Imperial College London.

XTRON built the XBOT-AEGIS architecture as a three-layer decoupled design. The top layer handles task understanding and decision making, the middle layer plans and simulates actions, and the bottom layer is a safety net that can intercept an abnormal instruction from the large model in milliseconds. The scheme balances the decision layer’s slow reasoning against the control layer’s fast response.

On the training side, XTRON is doing three things: pulling field data from deployed units and customer operation logs, building an internal simulation platform to supplement real samples with virtual data, and rolling out a VLA execution model while testing on the open-source JEPA world model as groundwork for its own foundation model.

XTRON XBOT-AEGIS three-layer embodied intelligence architecture diagram
The XBOT-AEGIS three-layer architecture splits decision, planning and safety interception. (Source: OFweek)

On that architecture XTRON plans a standardised line of explosive-proof embodied robots spanning inspection, manipulation and air-ground coordination, including wheeled and rail-mounted units, combined inspection-operation robots, air-ground units and explosive-proof biped wheel-foot robots.

Since 2025, its products have been delivered in multiple units to leading customers such as CNPC and Wanhua Chemical and have run stably. Overseas, the company pairs state-backed export with overseas distributors and has won project orders in South Sudan and Malaysia.

XTRON explosive-proof robot deployed at an overseas industrial site
An XTRON unit deployed outside China as part of its overseas order wins. (Source: OFweek)

The funding pace is quick. XTRON took a tens-of-millions Pre-A round from Oriental Fortune Capital in August 2025, then a Pre-A+ round in February 2026 with Qianhai Ark, Hangzhou Runmiao, Xihu and the Futian government fund. The A1 round came only half a year later, with Qianhai Ark and Xihu re-upping and industrial investor Guangyang coming in for the first time.

In hazardous settings XTRON has built a closed loop of technology validation, benchmark delivery and overseas orders. High safety barriers, rigid demand for labour substitution and strong paying customers mark what it calls the hard-but-right track within embodied intelligence.

Against general-purpose humanoids, XTRON has dug moats in explosive-proof compliance, accumulated industry know-how and earned customer trust. Continuous fast funding and industrial backing signal market confidence in its technical route and commercial loop, though long order cycles and high customisation will test its ability to hold delivery quality while scaling.

Editor’s note: This is an adapted translation of the original OFweek report. It has been trimmed and restructured for readability for an international business audience. The full original (in Chinese) is at https://robot.ofweek.com/2026-09/ART-8321204-12005-30701206.html.

Leave a comment