Chinese Robotics Firms Shift Focus to AI as Embodied Intelligence Attracts $13 Billion

Chinese robotics companies are shifting their attention from building humanoid machines capable of running, dancing and performing acrobatics towards developing the artificial intelligence needed to make robots useful in factories, warehouses and other workplaces.

The change is driving a surge in investment in embodied AI, a technology that enables robots to understand their surroundings, learn unfamiliar tasks and adapt to changing conditions without requiring programmers to provide instructions for every operation.

Chinese embodied AI companies attracted 93.5 billion yuan ($13 billion) across 322 investment deals in the first half of 2026, five times the funding recorded during the same period a year earlier, according to research firm ITjuzi.

Investment is increasingly flowing towards specialist developers of robotic intelligence, alongside companies manufacturing the machines themselves. The trend reflects growing recognition that advanced software, rather than hardware alone, will determine how useful robots become in commercial and industrial settings.

As competition pushes down the cost of robotic hardware, manufacturers are looking for new sources of revenue and technological differentiation.

Robot makers have increasingly concluded that future profits will come from developing more capable AI systems rather than simply producing machines, said Dermot McGrath, founder of Shanghai-based technology consultancy ZenGen Labs.

“The brain is indeed the weakest link in the complete robotics stack,” said Gao Yang, co-founder of Spirit AI, highlighting the technical challenge facing the industry.

The investment boom is also opening another front in the technology competition between China and the United States. American companies currently lead in some advanced robotic AI systems, while Chinese developers are seeking to challenge that position by drawing on the country’s extensive manufacturing capabilities and access to real-world training data.

Reuters interviews with founders and senior executives at five Chinese robotics companies, alongside industry analysts, indicate that businesses are increasingly concentrating on the software required to turn experimental machines into practical tools.

Embodied AI differs from conventional AI applications because it must connect digital intelligence with physical actions. Robots need to interpret visual information, recognise objects, understand their environment and respond appropriately when conditions change. They must also learn to carry out tasks that may not have been included in their original programming.

Developing these capabilities is considered essential if humanoid robots are to move beyond demonstrations and perform a wider range of jobs reliably.

The technology could have significant implications for manufacturing and logistics, where businesses are seeking greater automation and improved productivity. Robots capable of learning different tasks could potentially be deployed across multiple operations without extensive reprogramming.

The country that succeeds in combining advanced AI with physical machines could gain a competitive advantage in industrial automation and influence the future global market for robotics.

China’s growing investment in embodied AI suggests that the next stage of competition will depend increasingly on the intelligence powering robots, rather than their ability to perform impressive physical demonstrations.