XPENG Introduces XPACE to Enhance IRON Humanoid Learning with Human Videos
- Published
- Source
- HumanoidsDaily
XPENG Robotics has unveiled XPACE, a novel system designed to enhance the learning capabilities of its IRON humanoid robot. By utilizing 5,000 hours of human activity videos and simulated mistakes, XPACE aims to improve robot performance without the need for direct human demonstrations in every scenario. The system's innovative approach addresses a critical challenge in robotics training.
The significance of this development lies in its potential to advance humanoid robotics by enabling robots to learn from a broader range of experiences. XPENG's XPACE achieved an average success rate of 68.3% in tasks such as banana placement and water pouring, outperforming competitors like DreamZero and GR00T. This progress indicates a promising direction for the future of autonomous learning in robotics.
Looking ahead, XPENG's research highlights the importance of continuous improvement in robotic training methodologies. While the results are encouraging, the authors note that further optimization and testing are necessary to fully understand the benefits of the XPACE system. No further timeline was disclosed at the time of publication.
Original report
XPENG’s XPACE Trains IRON With Human Videos and Simulated Mistakes
HumanoidsDaily · [email protected] (Humanoids Daily Staff)This briefing is an independently written summary based on publicly available reporting and is provided for industry information and news discovery. The original report and source publication are credited and linked where applicable. RobotToday does not claim ownership of third-party source material.
Rights concerns? Contact [email protected] with the relevant URL and details. We will review the matter and take appropriate action where warranted.
More related news
- 1
-
2
Who Really Buys Humanoid Robots? Mostly Not Factories RobotToday · 20 hours ago
- 3
- 4
- 5
- 6
-
7
FAIR plus 2027 – Shenzhen, Apr 21–23, 2027 RobotToday · 2 days ago
- 8
- 9
-
10
RobotToday Weekly October 26 – 30, 2026 RobotToday · 3 days ago
- 11
- 12
Related Suppliers
Urban Drones
United States
South Florida drone retailer and distributor founded 2011, specialising in waterproof and underwater drones including Splash Drone, QYSea, and Chasing brands.
CIMON Inc.
United States
Henderson, Nevada and South Korea industrial automation maker founded 1999; builds PLCs, HMIs, industrial PCs, and SCADA software in house.
Beijing Ironman Technology Co., Ltd. 北京钢铁侠科技有限公司
China
Beijing developer (est. 2015) of the ARTROBOT series of large biped humanoids, now in its fifth generation, plus a 'motion brain' controller and core parts.
Qianxing Intelligence 深圳市迁星智能科技有限公司
China
Shenzhen maker, founded 2024, of player-controlled boxing mech robots for arcades and malls; its MECH ARENA concept runs in 30+ venues.
XPENG Inc. 小鹏汽车
China
Chinese EV maker expanding into autonomous vehicles and humanoid robotics; developing IRON humanoid robot targeting mass production in 2026.
Shuanglin Co., Ltd. 双林股份有限公司
China
Ningbo auto-parts maker (SZSE: 300100) mass-producing reverse planetary roller screws and linear joint modules for humanoid robots, incl. XPeng IRON.
Hubei Humanoid Robot Innovation Center 湖北人形机器人创新中心
China
Wuhan public platform opened in June 2025 to train humanoid robots and collect embodied-AI data in 23 simulated scenes; led by a Hubei state tech investor.
Hebbian Robotics
United States
San Francisco YC S26 start-up founded 2026 that builds HFlow, an open-source SDK for checking and curating robot training data.
Cortex Robot, Inc.
United States
YC F25 startup collecting egocentric workplace video and mobile bimanual robot trajectories, and building Cortex Harness for human-guided robot rollouts.
Zeroth 苏州乐享智能科技有限公司
China
Suzhou maker of Zeroth consumer robots, founded 2024: M1 home humanoid, W1 tracked carrier, Jupiter humanoid; over 30,000 units on order in 2026.
The Robot Learning Company
United States
YC X25 startup building open-source AI-native robot arms; TRLC-DK1 bimanual dev kit enables imitation learning and teleoperation for around USD 1,200.
Assured Robot Intelligence
United States
San Diego robot-AI start-up founded by Lerrel Pinto and Xiaolong Wang, building humanoid foundation models; acquired by Meta in May 2026.
Robotics needs a service framework.
RSF defines a common language for robot service capability, lifecycle operations, certification pathways, and service-provider networks.
Join the RobotToday community on LinkedIn
Daily robotics news, in-depth analysis, conference highlights, and discussions with professionals worldwide.
