NEURA Robotics launches Neuron AI, a robot model that adds touch, sound and memory to vision
NEURA Robotics said in a press release on Oct 14 that Neuron AI, its multimodal foundation model for physical AI, combines 3D vision, touch and force, audio, robot state and language with a memory function in what it calls a Perception-Language-Action architecture, so robots can act on contact and sound cues and keep track of objects out of view. The model was developed and trained in-house in Metzingen on real robot data from NEURA Gyms, simulation-amplified data and human demonstrations captured with sensor suits, runs across five NEURA robot embodiments, and draws on data partnerships with Schaeffler and Bosch; customers fine-tune it and deploy skills to fleets through the Neuraverse platform. NEURA said its 4NE1 humanoid will run an eleven-minute laundry sequence without human help at NVIDIA GTC Berlin, Oct 20 to 22. MEKAWIRE's records show NEURA's last round with a disclosed amount was a EUR 120 million Series B in January 2025, and the launch came a day after Google DeepMind released Gemini 4 Argon for robotics.