Digit’s Legs Make More Sense Once You Trace the Lineage
Digit’s unusual leg design makes more sense once you trace the line back through ATRIAS, Cassie, and the locomotion research that came before it.
One of the more interesting design stories in humanoid-adjacent robotics is Agility Robotics and Digit. A clear differentiator is the leg geometry. Instead of optimizing for a more human-like silhouette, the team ended up with a backward-bending, bird-like configuration that looks unusual at first but makes more sense once you trace where it came from.
That story goes back to Oregon State University and ATRIAS, a research biped developed to reproduce key human walking gait dynamics. From there came Cassie, which took the locomotion work further into a more practical legged platform before Digit ever existed as a commercial robot. OSU explicitly describes the sequence as ATRIAS, then Cassie, then Digit.
The shape of Cassie’s legs did not come out of nowhere either. OSU said the configuration emerged from research into the mechanics of legged locomotion and from solving inefficiencies in earlier designs, with Jonathan Hurst noting that the result looked like an ostrich or other ground-running bird even though the team was not trying to mimic animal appearance directly. They were trying to understand what made movement agile, efficient, and robust.
Cassie then gave the team years of real-world evidence before Digit. In 2021, it completed an untethered 5K on Oregon State’s campus on a single battery charge. In 2022, it set the Guinness World Record for the fastest 100 meters by a bipedal robot, running the distance in 24.73 seconds from a standing start and returning to a standing position without falling.
Digit’s final form looks less like a stylistic choice and more like the outcome of a long optimization process. Agility’s broader argument has always been that useful robots need to move through spaces built for people, including curbs, stairs, and existing facilities, without requiring the world to be redesigned around them. In that context, Digit’s legs look less strange and more like a practical answer to a specific locomotion problem.
So the interesting part is not just that Digit looks different. It is that the geometry points back to a research lineage that prioritized movement mechanics over human resemblance. That is a useful reminder that in robotics, the form that wins is not always the one that looks most familiar