We are an applied AI lab building LLMs that control robots.

We believe that the bitter lesson applies to robotics. Our thesis is that generalist models like LLMs will perform better than robot-specific models like VLAs. This means that we are counter-positioned against companies like Physical Intelligence.

We've raised $19M in seed funding, with participation from researchers at OpenAI, Google Deepmind, and Cognition. The founders met during RSI at MIT and studied math/physics at Harvard before leaving to start Waddle Labs.

Our mission is to define how humans work with robots. We're looking for early collaborators to help shape this vision.