Google has taken a major leap forward in robotics and artificial intelligence with the launch of its new Gemini Robotics On-Device AI model, which runs directly on robots without needing an internet connection. Unveiled by Google’s DeepMind division on June 24, this model marks a significant milestone in the development of autonomous machines capable of performing complex tasks in real time.
Building on the original Gemini Robotics model introduced in March 2024, the On-Device version has been specifically optimized to operate efficiently on robotic hardware. It can interpret natural language instructions much like ChatGPT and translate those into physical actions, enabling robots to carry out tasks with impressive general-purpose dexterity. Notably, the model is capable of handling previously unseen objects and environments, showcasing remarkable adaptability.
One of the standout features of Gemini Robotics On-Device is its ability to function offline. This makes it highly valuable for latency-sensitive operations where speed is critical and in areas with limited or no internet connectivity such as disaster zones, remote industrial sites, or space missions. The model processes data locally on the device, reducing dependence on cloud infrastructure and enhancing real-time responsiveness.
In a video demonstration shared by Google, the robot powered by Gemini On-Device successfully performs various everyday tasks, such as picking up objects it has never encountered before, sorting items, and reacting appropriately to dynamic environments all based on simple spoken instructions.
Google emphasizes that this new model is a step toward creating more autonomous, general-purpose robots that can safely and effectively assist humans in a wide range of settings, from homes to warehouses and hospitals. By running the AI model directly on the robot, Google avoids data privacy concerns and reduces latency, making it not only more efficient but also more secure.
As competition in the robotics AI space intensifies, Google’s Gemini Robotics On-Device model sets a new standard in offline, real-time robot control. With future updates, this technology could pave the way for smarter personal assistants and more resilient automation systems across industries.
