A humanoid robot developed by Galbot and Tsinghua University in Beijing's Haidian district has demonstrated real-time tennis rally capabilities, as the area accelerates the development and deployment of embodied artificial intelligence technologies.

The robot can track incoming balls, predict their trajectory, adjust its position and return shots through autonomous learning, whole-body coordination and real-time decision-making capabilities. It can also sustain multi-shot rallies against players of different ages and styles.

Its forehand return success rate reached 90.9 percent, highlighting progress in one of the toughest tests for humanoid robots, which must operate in dynamic, adversarial and unstructured environments, Beijing Daily reported.

Part of that progress comes from LATENT, a motion-learning system that enables the robot to learn from fragmented human motion data rather than complete professional datasets. It allows the robot to piece together skills such as ball perception, trajectory prediction, repositioning and shot execution into one continuous loop.

At a recent symposium held at ShanghaiTech University, Wang He, Galbot founder and chief technology officer, said he sees potential for legged humanoids in entertainment, education and sports training, including tennis practice scenarios. He added that the company hopes the robot will enter the sports market as an athletic training partner, according to a report by The Paper.

Galbot has been accelerating its commercial deployment. It has rolled out more than 100 robot-powered, round-the-clock instant retail stores in dozens of cities, processing more than 300,000 orders and selling about 1 million items to date, Workers' Daily reported.

Galbot's move reflects a broader industry trend as embodied AI expands across the country. Beijing's Haidian district, where Galbot is based, is strengthening its position in the sector.

The district has attracted more than 300 embodied-intelligence robotics innovators, including 24 humanoid robot manufacturers, three listed firms, five unicorns and 19 specialized "little giant" companies. It also has 21 universities and research institutes offering related programs, with a talent pool of about 19,000 people.

Tang Chao, deputy director of the administrative committee of Zhongguancun Science City and deputy district governor of Haidian, said the district is building a"1+X+1" modern industrial system. In that framework, AI serves as the lead engine, technology services as the support base, and strategic emerging as the central growth layer.

Embodied intelligence, she said, is one of the key priorities within that middle category. Haidian has also launched a 10 billion yuan ($1.46 billion) fund to guide social capital toward early-stage, small-scale, long-term and hard technology investments, building an integrated ecosystem of investment, industry and services to support the sector, Tang added.

Companies in Haidian said the district's broader coordination efforts are helping them reach users more directly and test services in real-world settings.

"Local authorities are helping firms connect with communities, subdistricts and end-users, while cooperation with universities is helping turn innovation into a closed-loop service system," said Dai Zhongyuan, director of elderly care products at Beijing Robint Technology Co.