The Future of Humanoid Robotics is Attritable

Cheap is critical. Exquisite is death.

Big news yesterday with NVIDIA’s announcement of Project GR00T, an ambitious effort to develop a foundation model for humanoid robotics. In layman’s terms, this project’s goal is to develop a universal standard to enable different types of humanoids (like 1X, Figure, Agility, etc.) to use a common base model for performing tasks.

This is very cool, but is only part of the solution. Despite the noise (and hefty capital raises) around humanoid robotics, the real commercial winners are likely to be those who make (dirt) cheap, virtually attritable humanoids that serve niche, critical, high value use cases. Things like firefighting, nuclear waste remediation, cell tower maintenance, and high risk military applications. Owning the customer will be key, something we’re already seeing play out in AI as incumbents use this new tech to entrench their advantage.

(Star Wars’ Battle Droid in The Phantom Menace)

In our view, delivering value at an extremely low capital cost is a technical and product positioning challenge that requires ingenuity, fast iteration cycles, and product constraints, not massive amounts of capital. It’s far easier to overfund complexity and ambitious research than make hard design and strategy decisions in service of simplicity and clear value delivery. From our vantage point, the capital flowing to the sector over the last 18 months is reminiscent of the famous Mark Twain line, “If I had more time, I would have written a shorter letter.” These companies will certainly develop new technology that advances the sector broadly, but it’s unclear where the key commercial products and services will come and the cost they’ll be able to provide those services. It’ll be interesting to see the rubber meet the road in the months and years to come!