The Tech Review 2025: China sees breakthroughs in AI and robotics

The year 2025 has been defined by a wave of transformative advances in China's artificial intelligence (AI) and robotics sectors, marking a decisive shift from laboratory research toward deep industrial integration and real-world application.

The year opened with a major shake-up in the foundational AI model landscape. DeepSeek, a company specializing in large language models, released a next-generation model that emphasized stronger reasoning and coding abilities rather than sheer parameter scale, while sharply reducing its application programming interface (API) costs. Widely described as a "game-changer," the move sparked global developer interest in usability and affordability, accelerating the rise of AI-native applications.

Momentum continued with growing public fascination around embodied AI, where intelligence is embedded in physical form. On January 16, humanoid robots developed by Unitree Robotics performed a highly synchronized dance at the China Media Group Spring Festival Gala. Showcasing dynamic balance, precise control, and swarm coordination, the performance reached hundreds of millions of viewers and became a cultural moment that brought advanced robotics into the public imagination.

Soon after, the physical limits of humanoid robots were put to the test. In April, Beijing hosted the world's first humanoid robot half-marathon. Tiangong Ultra claimed victory by completing the 21.0975-kilometer race in 2 hours, 40 minutes and 42 seconds. More than a contest of speed, the event served as a rigorous trial of endurance, hardware reliability, and real-world adaptability, key capabilities for future deployment in logistics, inspection and emergency rescue.

Alongside rapid technological progress, China also moved to strengthen AI governance. On September 15, the "Artificial Intelligence Safety Governance Framework 2.0" was released during the National Cybersecurity Awareness Week. Updated from its 2024 edition and developed under the guidance of the Cyberspace Administration of China, the framework incorporates expert and industry input to address emerging risks. Authorities say it aims to refine risk management, foster a secure and controllable AI ecosystem and promote international cooperation so that the benefits of AI can be shared globally.

"China has applied it very widely and on many different things. And in some things the west is catching up," Michael Levitt, 2013 Nobel laureate in chemistry, told CGTN in an interview in November.

"I think the robots will be very important bringing AI into the real world. I think China will be a leader," he said.

In sum, 2025 has drawn a clear picture of China's technological trajectory: powerful and increasingly accessible AI models serve as the "brain," while more capable and reliable robots act as the "body." Together, supported by strategic policy frameworks, they are driving China toward an intelligent, industrial future.