You have seen the Boston Dynamics videos. Atlas doing backflips. Spot dancing. Millions of views. Billions of dollars in investment.
You have not heard of Sangbae Kim. He runs the Biomimetic Robotics Lab at MIT. His robots do not dance for cameras. They do something more important: they exist so that other people can build robots too.
What he did
Kim designed the MIT Cheetah series: Cheetah 1, 2, 3, and the Mini Cheetah. These were among the first legged robots to run, jump, and do backflips autonomously without hydraulics. Using electric actuators instead, which is the critical detail.
Hydraulic robots are powerful but expensive, heavy, and impractical for anyone who is not a well-funded defense contractor. Electric actuators are lighter, cheaper, and accessible to research labs and startups.
Then Kim did something Boston Dynamics never did: he open-sourced much of the Mini Cheetah design. The CAD files. The control algorithms. The documentation. Free.
What happened next
The affordable quadruped robot market exists because of that decision. Unitree (the company making robots that play tennis and cost a fraction of Spot) traces significant design DNA back to Kim’s open-source work. Hundreds of university research labs worldwide are using Mini Cheetah derivatives.
When you see a quadruped robot that costs $2,000 instead of $200,000, the price difference is largely because one professor decided that legged robotics should not be locked behind corporate walls.
The recognition equation
Boston Dynamics gets the YouTube views. They have a marketing team, a PR strategy, and a parent company (Hyundai) that needs to justify a massive acquisition. Their robots are impressive. Their videos are engineered to go viral.
Kim gets cited in academic papers. His students go on to found robotics companies. His designs spread through the ecosystem invisibly, embedded in products that never mention his name.
That is how foundational contributions work. The person who opens the door rarely gets credit for what walks through it.
Why this matters now
Physical AI is the next frontier. Humanoid robots on factory floors. Autonomous vehicles on roads. Drones in warehouses. All of them need to move through physical space, and all of them build on decades of research into how machines interact with the real world.
Kim’s contribution was not just a robot. It was the idea that the research should be shared. That making legged robotics accessible to everyone would accelerate the field faster than keeping it proprietary.
He was right. The proof is in every affordable robot that exists today.
Professor Sangbae Kim leads the Biomimetic Robotics Lab at MIT.