Teaching robots no longer requires coding or remote control. All you have to do is show them and tell them.
Israeli-based startup Mentee Robotics has released a video showing its Menteebot responding to a request to retrieve a can of cola, performing the task smoothly and autonomously.
The video also shows how the robot interprets commands, plans its next actions, and completes its tasks, with footage from its onboard camera.
On the same day, Israeli self-driving company Mobileye announced it would acquire Mentee in a bid to expand its self-driving cars and humanoid robots globally.
Teach robots naturally
In the shared video, when asked to bring a can, the Menteebot V3 humanoid first interprets the vocal request and then scans its surroundings to identify the correct item. It autonomously moves to the counter, finds and grabs the matching can, and returns to the person.
Mentee said the demonstration highlights how language understanding, visual recognition, navigation, and object manipulation are seamlessly integrated into a single, coordinated movement. “This approach significantly lowers the barriers to human-robot collaboration and enables non-experts to effectively deploy humanoid robots in real-world environments,” the video description reads.
Similar to new employee onboarding, Mentee uses a “foundational model” that allows humans to train robots through verbal instructions and live demonstrations, allowing robots to learn naturally without complex programming.
Training begins in a simulated environment using reinforcement learning, and the robot improves through trial and error. Skills developed in simulation are transferred to real-world machines using Sim2Real techniques to help reduce risk and speed deployment.
This model combines two core methods: learning from experience, which focuses on controlling actions and tasks, and learning from data, which includes verbal, visual, and audio input. According to Mentee, this blend allows robots to learn faster and adapt more flexibly.
MenteeBot is designed to help you with everyday chores like setting the table and doing laundry, and to quickly learn new tasks through verbal instructions and visual imitation. In warehouses, it supports handling, transporting and organizing items, lifts up to 25 kilograms and operates for up to 3 hours on a single charge.
Augmenting physical AI
Founded in 2022, Mentee Robotics is chaired by Mobileye founder Professor Amnon Shashua, and co-founded by CEO Professor Lior Wolf (former senior scientist at Facebook AI Research) and machine learning expert Professor Shai Shareef Schwartz. The company has raised more than $40 million and has approximately 70 employees.
At CES 2026 in Las Vegas on January 6, Mobileye revealed plans to acquire Mentee in a deal worth up to $900 million. While the immediate benefits for Mobileye may not be obvious, the company is positioning the move as a strategic expansion beyond automotive technology, according to TechCrunch.
The companies say self-driving cars and humanoid robots face common challenges in functioning safely and efficiently in environments designed for humans. Both must meet stringent performance standards, run on efficient edge computing systems, provide provable security, and scale in a commercially viable manner. As a result, the two fields rely on the same physical artificial intelligence foundation, combining multimodal perception, world modeling, intention-aware planning, precise control, and decision-making under uncertainty.
Mobileye's acquisition of Mentee is expected to unlock strategic synergies across this shared technology stack. Mentee's advances in visual-verbal-behavioral models, large-scale simulation, and advanced Sim2Real transfer directly complement Mobileye's autonomy platform, improving generalization, accelerating development, and enabling faster adaptation to new environments.
The companies say Mobileye is implementing a safety-first framework honed in autonomous driving, including formal models such as responsible safety and mathematically-based decision-making. Applying these tools to humanoid robots supports predictable and auditable behavior in dynamic human spaces. Combined with Mobileye's commercialization expertise, this partnership will accelerate the safe, large-scale deployment of humanoid robots in industrial settings around the world.
