Boston Dynamics and NVIDIA are joining forces to shape the future of humanoid robots. Both companies are leaders in their fields—Boston Dynamics in advanced mobile robotics, and NVIDIA in artificial intelligence computing. Their renewed partnership stands as a major step toward creating robots that are not only strong and agile, but intelligent and aware.
The Power Behind Atlas
At the heart of this effort is Atlas, Boston Dynamics’ most advanced humanoid robot. Atlas can run, jump, and handle a variety of objects. But to make Atlas smarter, Boston Dynamics is now relying on NVIDIA’s latest technology.
Central to this is the NVIDIA Jetson Thor computing platform. Jetson Thor is a powerful, small computer built specially for robots. It can run complicated AI programs without slowing down—giving Atlas the brains it needs to do tough, detailed tasks. This platform lets Atlas coordinate its whole body for smooth, precise motion and manipulation, even in difficult and dynamic settings.
Boston Dynamics is also one of the first companies to use NVIDIA’s Isaac GR00T framework. This tool, along with NVIDIA’s Isaac Lab and simulation platforms like Omniverse and Isaac Sim, creates a virtual playground for robots. Inside these simulated worlds, Atlas can practice new skills thousands of times before trying them in real life. This is key for “learned dexterity”—the robot’s ability to handle delicate and unpredictable situations with its hands and body.
Shared Vision: Smarter, Safer Robots
The deeper collaboration has clear goals. The two companies want to demonstrate just how intelligent humanoid robots can become by combining superior hardware with the latest in AI.
Main objectives include:
- Learned Dexterity and Locomotion: Atlas should be able to move smoothly across complex areas—walking, running, stepping over obstacles—and use its hands for challenging tasks.
- Functional Safety and Security: As robots become more capable, keeping them safe and secure is critical. The partnership will set tough standards for reliability and trust, whether Atlas is working with people or on its own.
- Advanced Computer Vision: Atlas will use state-of-the-art “vision pipelines” to see, recognize, and understand its environment. This means it can better interpret what’s around it and respond in real time.
Bridging Simulation and the Real World
A milestone in this collaboration is the connection between what robots learn in virtual spaces and how they perform in reality. As Aaron Saunders, Chief Technology Officer at Boston Dynamics, puts it: “Robots are the bridge between simulation and the real world.” Through this partnership, the lessons Atlas learns in NVIDIA’s simulations quickly translate into real-world skills.
This approach lowers the risks and costs of training robots while speeding up the journey from lab to life. If Atlas can master a skill in simulation, it’s more likely to succeed at it when interacting with people, machines, or challenging environments.
Looking Ahead
Boston Dynamics and NVIDIA are not limiting these advances to Atlas alone. The same cutting-edge AI is being adapted to other robots like Spot, the dog-like machine famous for its agility and perception, and to systems like Orbit, which manages fleets of robots working together.
By joining their strengths, Boston Dynamics and NVIDIA are setting new standards for what robots can do. The hope is to create intelligent machines that are not only more flexible and capable, but also safer and more responsive to the world around them.
This collaboration is a turning point in robotics. As technology progresses, humanoid robots like Atlas may soon contribute more widely—from assisting in factories and warehouses to playing a role in search-and-rescue, healthcare, and beyond. The work done today lays the foundation for a future where robots are trusted partners, serving society with intelligence and care.
Leave a Reply