German company Neura Robotics has launched the Neuron AI foundation model for physical artificial intelligence, combining 3D vision, touch, sound, and memory. The model will undergo its first public test next week in Berlin when the 4NE1 humanoid robot performs laundry tasks autonomously.
Launching Neuron AI to build a unified brain for physical robots
Perception-language-action architecture and local cloud-free operation
Claims awaiting field testing at the Berlin station
A well-funded German competitor with a valuation approaching seven billion dollars
Frequently asked questions
Launching Neuron AI to build a unified brain for physical robots
German robotics company Neura Robotics, headquartered in Metzingen, unveiled its new “Neuron AI” model on Wednesday, October 14, 2026. The company describes this innovation as a multimodal foundation model dedicated to the field of “physical artificial intelligence,” combining 3D vision, touch, sound, robot motor state, natural language, and continuous memory functions. This allows robots to act on information that cameras alone cannot capture and to continue tracking objects even after they leave the field of view.
This launch represents the German company’s bet on building a unified AI “brain” that works across its entire hardware and automated equipment product line.
The first public test of this model will take place next week when the company’s humanoid robot, “4NE1,” attempts to execute a complete 11-minute laundry task during the Nvidia GTC Berlin conference, where Neura asserts the robot will complete the task entirely without human assistance or intervention.
Perception-language-action architecture and local cloud-free operation
Neura refers to this engineering design as the “perception-language-action” architecture—a software framework linking what the robot sees, hears, and feels with the actions it has previously executed. The model places all these sensory inputs into a single shared space over time, transforming them into a coherent stream of motor actions. The company provides practical examples of this importance, noting that sound and touch alone can confirm the opening of a mechanical lock, and the sound of an industrial welding process can reveal the quality and precision of a weld joint.
David Reger, founder and CEO of Neura Robotics, stated: “An intelligent robot needs a brain that combines its senses together and remembers what happened. With Neuron AI, we are building the brain of the physical world.”
According to the company’s product page, the model runs directly on the robot’s own processors rather than relying on remote servers, enabling it to continue working without a cloud connection and keeping industrial production data under the customer’s full control. Model training relies on three main sources: first, real robot data from the company’s training sites known as “game rooms”; second, digital simulations of tasks; and third, recordings of people wearing suits equipped with motion sensors. The model operates across 5 types of Neura robots, and customers can adjust and deploy new skills across their fleets via the “Neuraverse” platform, with the company adopting “time to skill acquisition” as a key metric for progress, supported by data partnerships with Schaeffler and Bosch.
Claims awaiting field testing at the Berlin station
Despite the grand ambitions surrounding the announcement, Trending Topics reported that Neura Robotics has not yet published official benchmark test results for the Neuron AI model, meaning independent experts have not yet been able to compare its performance claims with competing robot models on the market.
To compensate with a live operational demonstration, the humanoid robot “4NE1” is scheduled during GTC Berlin to sort dirty laundry, place it inside a washing machine, close the door and reopen it after the cycle finishes, and then unload the washing machine and transfer the clothes to an electric dryer.
Visitors and specialists will be able to watch this live demonstration at the STATION-Berlin exhibition center from October 20 to 22 to judge the smoothness and independence of the new model.
A well-funded German competitor with a valuation approaching seven billion dollars
The launch of the Neuron AI model follows Neura Robotics’ announcement in June of a Series C funding round reaching $1.4 billion USD, making it one of the largest funding rounds in the European robotics sector.
The list of investors supporting the company in that round included major technology giants and financial institutions such as Nvidia, Amazon, Qualcomm, Tether, Bosch, and the European Investment Bank.
CNBC reported, citing a person familiar with the matter, that the company’s valuation reached approximately $7 billion, explaining that obtaining the full amount of the funding round depends on the company achieving specific milestone targets. Neura confirmed that the value of its current purchase orders and pending deployment contract pipeline already exceeds the one billion dollar mark.
Frequently asked questions
Question: What senses and capabilities does the new Neuron AI model combine?
Answer: The model combines 3D vision, touch, sound, robot state, and language, with continuous memory within a perception-language-action architecture.
Question: Does the Neuron AI model require a cloud connection to operate robots?
Answer: No, the model runs directly on the robot’s own hardware without needing a cloud connection, preserving the privacy of customers’ industrial production data.
Question: What is the public test that the 4NE1 robot will undergo in Berlin?
Answer: It will execute a complete 11-minute laundry task without human assistance during the Nvidia GTC Berlin conference between October 20 and 22, including sorting, operating, and transferring to the dryer.
Question: What is the funding size and financial valuation of German company Neura Robotics?
Answer: The company announced a funding round of up to $1.4 billion in June with a valuation approaching $7 billion, supported by Nvidia, Amazon, Qualcomm, and Bosch, with orders exceeding $1 billion.
