The new Google AI makes the most intelligent robots without the cloud

Google Deepmind has introduced a powerful version in the device of its Robotics Gemini.

This new system allows robots to complete complex tasks without trusting a cloud connection. Known as Gemini Robotics on the device, the model carries the advanced reasoning capabilities and gemini control directly to physical robots. It is designed for rapid and reliable performance in places with poor or zero Internet connectivity, which makes it ideal for real -world latency environments.

Google working to decode the communication of the dolphins using AI

Smarter robots that work anywhere

Unlike its predecessor connected to the cloud, this version is fully executed in the robot itself. You can understand natural language, perform fine motor tasks and generalize from very few data, all without requiring an Internet connection. According to Carolina Parada, Robotics Chief on Google Deepmind, the system is “small and efficient enough” to operate directly on board. Developers can use the model in situations where connectivity is limited, without sacrificing intelligence or flexibility.

What is artificial intelligence (AI)?

Easy to adapt and train

Gemini Robotics On the device can be customized with only 50 to 100 demonstrations. The model was first trained using Google’s Robot Aloha, but has already adapted to other platforms such as APPTRONIK and Franka Fr3. For the first time, developers can adjust a deep robotics model. Google offers access through its trusted tester program and has published a complete SDK to support experimentation and development.

Local control means more privacy and reliability

Since artificial intelligence is executed directly in the robot, all data remain local. This approach offers better privacy for sensitive applications, as in medical care. It also allows robots to continue operating during Internet interruptions or in isolated environments. Google sees this version as a strong adjustment for remote configuration, sensitive to safety or poor infrastructure. The system offers faster response times and less failure points, opening new possibilities for the implementation of robots in real world configurations.

Security requires developer entry

The device model does not include incorporated semantic security features. Google recommends that developers create security systems in their robots using tools such as the API Live Gemini and low -level low level controllers. The company is limiting access to selected developers to better study security risks and real world applications. While the hybrid model still offers more general power, this version remains its own for more common cases and helps push robotics closer to daily implementation.

Kurt’s Key Takeways

The launch of Gemini Robotics on the device marks a turning point. Robots no longer need a constant cloud connection to be intelligent, adaptable and useful. With faster performance and stronger privacy, these systems are ready to address real world’s tasks in places where traditional robots could fail.

Click here to get the News application

Would you feel comfortable delivering tasks to a robot that does not need internet to think? Get us knowing in Cyberguy.com/contact.

Register for my free Cyberguy report
Get my best technological tips, urgent security alerts and exclusive offers delivered directly to your inbox. In addition, you will get instant access to my definitive scam survival guide, free when it joins me Cyberguy.com/newsletter.

Copyright 2025 Cyberguy.com. All rights reserved.