Pudu Robotics product portfolio
News

Pudu Robotics Debuts D7 Semi-Humanoid Robot

2 MINUTE READ|AI NewsAI News|Jul 19, 2026
Michelle Hawley avatar
By
SAVED
Pudu Robotics unveiled its D7 semi-humanoid robot at WAIC 2026, demonstrating industrial tasks and coordination with a quadruped robot.

Key Takeaways

  • Pudu Robotics debuted its D7 semi-humanoid robot at WAIC 2026.
  • The D7 is designed for manufacturing, warehousing and retail tasks.
  • Pudu is building a shared AI platform across multiple robot types.

Pudu Robotics publicly demonstrated its new PUDU D7 semi-humanoid robot at the World Artificial Intelligence Conference, held from July 17–20 in Shanghai. The company shared plans to expand from specialized service machines into more adaptable industrial robots.

During the conference, the robot autonomously walked through the company’s booth while guiding Pudu’s four-legged D5 robot and avoiding people in its path. Attendees could also pose for photographs taken from the robot’s perspective.

The demonstration aimed to show how robots with different physical designs could coordinate inside the same environment, the company said.

PUDU D7 Targets Industrial and Commercial Work

Pudu described the D7 as an industrial-grade semi-humanoid robot intended for manufacturing, warehousing and retail environments.

Unlike fully humanoid robots that replicate the entire human form, semi-humanoid systems generally combine humanlike manipulation capabilities with a body optimized for stability or a particular working environment.

According to a separate Pudu Robotics technology announcement, the D7 is designed to perform tasks such as:

  • Move materials inside industrial facilities
  • Pick items from elevated locations
  • Support warehouse logistics
  • Operate in manufacturing and retail settings

The robot supports payloads of up to 14 kilograms — or about 31 pounds — and can operate at heights of up to 2 meters, or approximately 6.6 feet.

Pudu did not announce pricing, customer availability or a commercial release date for the D7.

One AI System Across Multiple Robot Designs

The D7 is part of Pudu’s “One Brain, Multiple Embodiments” strategy, through which specialized, quadruped, semi-humanoid and humanoid robots share components of the same underlying AI and software architecture.

The company’s three-layer robotics platform includes:

LayerFunction
Robot bodyProvides different physical designs for specific environments and tasks
PuduAgent OSConnects AI models, robot skills, memory and safety controls
PuduFMHelps robots perceive environments, plan tasks and execute physical actions

Pudu said the shared architecture could allow capabilities developed for one robot type to be reused across other machines and operating environments.

The company is positioning specialized robots for repetitive and standardized work, semi-humanoid robots for more complicated commercial and industrial settings and humanoid robots for less structured collaboration with people.

Learning OpportunitiesView All

Pudu Expands Beyond Delivery and Cleaning Robots

Pudu also displayed several of its established commercial robots at WAIC, including its BellaBot delivery robot, T300 industrial delivery robot and MT1 Max and CC1 Pro cleaning robots.

Pudu Robotics' BellaBot
Pudu Robotics' BellaBot

Its D5 quadruped robot navigated an obstacle course containing sand, gravel, steps and slopes designed to recreate conditions found in locations such as industrial parks and power substations. Pudu said the D5 can reach speeds of up to 5 meters per second.

The company said it has shipped more than 130,000 robots across over 85 countries and regions. Its machines are deployed in hospitality, retail, manufacturing, warehousing, health care, real estate and other commercial environments.

Pudu also said it raised nearly $150 million in April, bringing its valuation above $1.5 billion and its cumulative funding to more than $300 million.

Editor's Note: For more on physical AI...

Main image: Pudu Robotics

About the Author

Michelle Hawley is Editorial Director at VKTR and host of The Inference. She covers the evolving AI landscape, including AI infrastructure, LLM development and enterprise AI strategy. With more than 10 years of experience, she has written for various publications, including The Press Enterprise and The Ladders, and taught courses on writing at Lycoming College.
Featured Research