The Spatial Motion research lab—foundation models to decode the dynamics of systems to unlock new human capability.

Only recently has human and animal motion become available at scale, with diversity and high authenticity through Move AI's markerless motion capture and hybrid solutions. Previously, motion had to be inferred from contrived environments—people and animals in labs, not in their natural environments, wearing restrictive suits, or simply from videos. Move AI, together with partners, enables large-scale capture of authentic movement in real-world settings.

We're learning from human and mammalian life through multi-view, rich information captured in natural environments—not just from language. This is the same information that a child learns from: egocentric experience and watching, wondering curiously about other life forms. Our models absorb the spatial intelligence that emerges from billions of real interactions, not synthetic approximations.

We believe that human and machine interaction with the world is fundamental to intelligence. True spatial understanding emerges from the dynamic interplay between intent, action, and consequence across diverse environments. By studying how living systems navigate, adapt, and thrive in the physical world, we're building foundation models that capture the essence of embodied intelligence.