Google’s new AI can now walk a humanoid across the room and tidy up, feet to fingertips

What changed

Google DeepMind launched Gemini Robotics 2, a new AI system that controls entire humanoid bodies—from feet to fingertips. This means one AI brain can coordinate every joint and limb, enabling a humanoid robot to walk across a room and interact with objects in a human-like way. The system also handles multiple robots simultaneously and adapts to different tasks and environments.

Why builders should care

Gemini Robotics 2 simplifies robot control by unifying movement and manipulation under a single AI model. Builders won’t need separate systems for walking, grasping, or balancing, which streamlines development and reduces integration headaches. The AI’s ability to coordinate multiple machines opens up new possibilities for warehouses, factories, or service settings where several humanoid robots might work together seamlessly.

The practical takeaway

For companies building humanoid or multi-robot systems, Gemini Robotics 2 promises faster deployment and more flexible robot behaviors. Instead of hard-coding motions for every scenario, this AI can generalize, potentially lowering development costs and speeding up adaptation to changing tasks. Operators could see robots that actively tidy spaces, assist with manual labor, or support interactive use cases without constant human reprogramming.

What to watch next

Look for real-world applications and demonstrations of Gemini Robotics 2 outside labs. The key will be how well this AI performs in unpredictable environments and if it handles fragile or complex tasks reliably. Also, watch how Google DeepMind integrates this tech into commercial or open platforms, which will reveal if competitors face higher pressure to unify robot controls or rethink their multi-agent strategies.

AI Quick Briefs Editorial Desk