- Founder argues adaptability, not terrain, is the biggest obstacle to mass adoption
- Reliability, autonomy and cost will determine whether robots scale beyond industrial use
The biggest challenge facing robots today is not navigating rugged outdoor terrain but operating inside ordinary homes, according to the founder of Chinese quadruped robot maker DEEP Robotics (云深处科技).
Zhu Qiuguo (朱秋国), chief executive of the Hangzhou-based startup and incumbent professor teaching at Zhejiang University, said the industry’s next breakthrough will depend on machines adapting to people rather than the other way around.
Speaking at a panel titled “When robots learn to be human” during the 17th Summer Davos forum in Dalian on June 23, Zhu said robots have already demonstrated strong performance in structured environments but continue to struggle when confronted with uncertainty.
Struggling in an uncertain world
“For deterministic tasks, robots are very capable. In an uncertain world, they are still very weak,” Zhu said.
The Summer Davos forum runs from June 23 to 25 in the northeastern coastal Chinese city of Dalian.
Zhu’s comments come as embodied AI and robotics attract growing investment globally, with developers racing to move machines beyond controlled industrial settings and into everyday environments.
Drawing on DEEP Robotics’ experience in deploying quadruped robots, Zhu said the company has accumulated more than 1,200 real-world application cases. Its robots have achieved inspection accuracy rates of 96.5% in substations and logged more than 1,000 hours of fault-free operation, he said.
Homes pose harder tests
Yet despite those achievements, he argued that domestic environments remain far more challenging than many industrial or outdoor settings.
Outdoor terrain is generally fixed and can be modeled in advance, he said. Homes, by contrast, are highly dynamic spaces filled with furniture, cables, slippers and pets, with no standardized layouts.
“That is the core barrier preventing service robots from achieving large-scale adoption,” Zhu said.
To overcome that hurdle, Zhu said robotics development needs to shift from designing environments around machines to designing machines around humans.
Making robots adapt to humans
“The logic of development must change from making scenarios adapt to robots to making robots adapt to humans,” he said, adding that robot body structures, locomotion, perception systems and operating logic should be aligned more closely with human environments and behaviors.
At the same time, Zhu cautioned that robots do not necessarily need to take humanoid form. The ultimate goal, he said, is to handle repetitive and hazardous tasks while freeing people for higher-value work.
He identified three prerequisites for broad adoption: reliability, autonomy and affordability.
“Reliability is the foundation, autonomy is the core, and cost is the prerequisite for scaling,” Zhu said. “The ultimate value of robots is not to replace people, but to augment, serve and liberate them.”
