Google DeepMind and robotics company Apptronik have announced a partnership to combine their technological expertise in developing humanoid robots that can operate in complex dynamic environments. This collaboration aims to promote the deep integration of robotic hardware and artificial intelligence, exploring broader possibilities for robots in industrial and everyday applications.
Founded in 2016 as a spin-off from the Human Centered Robotics Lab at the University of Texas at Austin, Apptronik has developed 15 robot models to date, including the Valkyrie robot designed for NASA. The company's latest release, the Apollo robot, stands 1.73 meters tall and weighs 73 kilograms, specifically designed to work alongside humans in industrial settings. Google DeepMind recently showcased the Apollo unit operating with its AI model Gemini 2.0.
As a leader in robotic hardware, Apptronik has established partnerships with GXO and Mercedes-Benz, and plans to expand its collaborations in the future. Google DeepMind, on the other hand, possesses top-tier expertise in machine learning and physical simulation, with its latest AI model Gemini 1.5 Pro demonstrating exceptional environmental navigation capabilities, achieving a 90% success rate in complex tasks. Additionally, this AI system can learn through simple smartphone videos, recognizing small objects like coasters and understanding abstract concepts, such as accurately locating a whiteboard area when searching for a suitable place to paint.
Currently, the competition between AI and robotics technology is intensifying. Google DeepMind's main competitor, OpenAI, has recently re-entered the robotics field through a partnership with Figure and is rebuilding its previously closed robotics division.
This collaboration between Google DeepMind and Apptronik not only injects cutting-edge AI capabilities into the Apollo robot but also paves the way for the practical application of humanoid robots in complex environments.
