Foundation models that enable digital humans to see, understand, and move.