
Chinese embodied AI startups raised 93.5 billion yuan in the first half as companies compete to make robots adapt to more tasks.
Chinese embodied AI companies raised 93.5 billion yuan ($13 billion) across 322 deals in the first half of 2026, five times the amount a year earlier, according to research firm ITjuzi. The investment is reaching software developers as well as robot manufacturers.
The industry is shifting its focus from machines that can run, dance or do backflips to AI that helps robots understand their surroundings and learn new jobs. Companies are seeking software that can adapt without being programmed separately for every task.
U.S. developers including Physical Intelligence, Skild AI and Generalist are working on models that can give robots new skills. In China, X Square, Galaxea and Spirit AI are focusing on developing systems with manufacturers and industrial customers in real work settings, Dünya reported.
Training requires examples of physical tasks, such as gripping objects or handling flexible cables. Generalist said its GEN-1.5 model achieved an average 59% success rate on 10 simple tasks after a single demonstration lasting three to 12 seconds. The company said the result applied to limited tasks, and skills learned from short demonstrations were less robust than those given extra task-specific training.
Some robots have reached workplaces, though use remains focused on specific jobs. Spirit AI said its robots perform battery-pack testing connections at CATL’s Zhongzhou facility, with success above 99% for those connections. Dünya also reported that Robbyant robots help prepare orders at two Shanghai pharmacies, where employees check medicines before delivery.
Access to advanced chips is another constraint for Chinese developers, according to Dünya. Robotics consultant Georg Stieler warned that reliable commercial use across unfamiliar tasks could take years after successful demonstrations.
This article was produced with the help of AI technology. Based on reporting by the outlets listed above.