← DeepTechTracker

NEURA introduces Neuron AI with multisensory robot control and memory

NEURA Robotics has introduced Neuron AI, a Physical AI foundation model combining 3D vision, touch and force data, audio, robot state and language with memory. The company says its Perception-Language-Action architecture is trained on real robot data, simulation-expanded data and human demonstrations across five NEURA robot embodiments. Customers can fine-tune skills and deploy them through Neuraverse. These are company-described capabilities, not independently benchmarked reliability or general-purpose autonomy. NEURA plans an autonomous 11-minute laundry-handling demonstration with its 4NE1 humanoid at NVIDIA GTC Berlin on October 20-22; the demonstration was still upcoming at the time of the October 14 announcement.

Read at source ↗

DTT summary based on NEURA Robotics / EU-Startups's report. The full article is on the publisher's site.

Inputs

3D vision, touch/force, audio, robot state and language

Architecture

Perception-Language-Action with memory, per company

Next demonstration

4NE1 laundry sequence planned for October 20-22

Official NEURA promotional image of robotic hands clipping a component, not independent proof of task reliability
Official promotional image: NEURA Robotics
Browse more deep tech news →