NVIDIA’s GROOT AI is giving robots unbelievable human-like moves

NVIDIA is stepping up its AI game with the launch of its open-source, pre-trained AI model for humanoid robots. Dubbed the Isaac GR00T N1 (or just Groot N1), this "generalist" model was revealed at the GTC 2025 conference in California on March 18.

During his keynote, NVIDIA CEO Jensen Huang made a bold statement: “The age of generalist robotics is here.” With this launch, NVIDIA is throwing its weight behind the future of humanoid robotics, making its AI tech publicly available to developers worldwide.

AI that thinks fast, and slow

Groot N1 was trained on a mix of real and synthetic data, and NVIDIA plans to release its simulation tools and training blueprints to the public. According to a video from NVIDIA, the model has a “dual-system architecture” inspired by human cognition, meaning it can process information in two ways—quick reflexive responses and slower, more deliberate reasoning. In simple terms, this AI isn’t just reacting; it’s thinking.

NVIDIA joins the race, but with a different play

Let’s be real—NVIDIA isn’t the first to jump into humanoid robotics. Companies like Tesla (with its Optimus robot) and Figure AI have been in the game for a while. But NVIDIA isn’t here to build robots from scratch. Instead, it's providing the AI brains, computing power, and software tools to help others build smarter, more capable robots.

To push things further, NVIDIA is teaming up with Boston Dynamics, the company behind the famous Atlas robot, to integrate its Jetson Thor AI computing platform into humanoid robots. This move could help accelerate real-world applications for these machines.

The billion-dollar bet on robots

The humanoid robot market is on track to hit $38 billion in the next decade, with demand skyrocketing in industries like manufacturing. NVIDIA sees factories as the first real proving ground for these robots—controlled environments with structured tasks that AI-powered machines can handle efficiently.

And the cost? Huang suggests that renting a humanoid robot could set businesses back about $100,000 per year, which could still be a bargain for companies looking to cut labor costs and improve productivity.

A tipping point for AI-powered robotics?

There are still plenty of hurdles to clear—robots need more training data, better energy efficiency, and smarter AI models. But Huang believes we’re at a “tipping point” in computing, and with NVIDIA’s AI muscle behind it, the humanoid robot revolution might be coming a lot sooner than expected.

NVIDIA may not have started this race, but it’s hitting the gas—and the future of humanoid robots just got a serious upgrade.

Image and video courtesy: NVIDIA