Every humanoid robot company has a founding story; Booster's happens to be unusually long. The company itself was founded in 2023, but its core team comes from the Tsinghua Vulcan robot football team - a research group with more than 20 years of work in humanoid robotics, competition results stretching back to 2006, and publications at the highest level of the field. The Booster story is, in large part, the story of what happens when a competition team decides its robots should leave the university and go to work.
From a University Team to a Company
The Vulcan team spent two decades doing the least forgiving kind of robotics research: building bipedal robots that had to perform in public, against opponents, on a clock. Robot football does not allow rehearsed demos. Across years of international competition, the team accumulated the things that cannot be bought quickly - an intuition for how humanoids fail, a control and perception stack refined over hundreds of matches, and a generation of engineers trained on real hardware. When Booster was founded in 2023, it inherited that stack and that instinct, and pointed them at a commercial question: what would laboratories, schools and developers buy, if the competition platform became a product?
Twenty Years of Competition Schooling
Competition is a brutal but honest teacher. It rewards robots that walk fast and recover faster, and it punishes fragile mechanics, brittle software and optimistic specifications without mercy. The Vulcan heritage is visible in what Booster prioritises. Fall-and-get-up is treated as a core skill, not an edge case, because in football a robot that cannot rise is a robot that has left the match. Disturbance rejection - staying upright when bumped - is trained deliberately. And the whole development loop is built for iteration speed, because a competition team that cannot test a fix before the next match does not stay a competition team for long.
Heritage You Can Inspect
Heritage claims are cheap in this industry, so Booster keeps its own inspectable. The timeline on the company page runs from the Vulcan team's early results in 2006 through the founding in 2023 to the 2025 world title, and the connecting thread is documented in competition records and publications rather than in advertising copy. Visitors to the robot football training base can watch the current generation of the same programme training on the same platform class customers buy. When a supplier tells you a founding story, ask what of it you can verify - results, papers, a facility, a team. In Booster's case, the answer is: nearly all of it.
What the Heritage Changes in the Product
Open the Booster T1 specification and the lineage shows. Its 23 degrees of freedom are paired with a reinforcement learning framework in which walking, kicking, dribbling, disturbance rejection and fall-and-get-up are developed, deployed and verified through one pipeline - the same workflow the founding team used competitively. The smaller Booster K1 carries the same philosophy into a 0.95 m classroom-friendly format. In both cases the design brief was written by people who had spent twenty years picking robots up off the floor: make it capable, make it recoverable, and make the software loop fast enough that users improve every week.
The Same Sideline, New Players
The clearest proof that the heritage transferred came in 2025, when a Chinese university team won the RoboCup world title - the first for China in 28 years - on a Booster-class platform, and Booster-powered entries swept the football medals at the World Humanoid Robot Games. The founding generation's knowledge is now compounding through hundreds of customer teams instead of one. That, ultimately, is what the Vulcan story means for a buyer: you are not purchasing a startup's first attempt at a humanoid. You are purchasing the product form of a twenty-year research programme, with the trophies - and the scars - to show for it.
Want the full timeline, from the Vulcan team's early results to the 2025 world title? Compare the relevant product range and deployment scenarios to take the next step.
