PhD-student founders bet on an all-in-one brain for bipedal humanoids as their robot drives a go-kart fully autonomously

A startup founded by several PhD students is drawing attention with an unconventional bet in the humanoid robot race: instead of following Silicon Valley, it is betting on an all-in-one brain for bipedal humanoids. In the team's latest demonstration, reported by QbitAI, a robot drives a go-kart through consecutive corners, fully autonomous in one continuous take. Go-kart driving is a demanding test for a robot, requiring dynamic balance and fast, continuous control decisions at speed with no human intervention. Unlike mainstream designs that separate perception, planning, and motion control into distinct modules, the integrated-brain approach fuses understanding, decision-making, and control into a single end-to-end system. The team chose the bipedal humanoid form as its carrier, betting on the future of general-purpose robots, and the demo is a test of whether the integrated brain holds up in a real physical scenario. If the route works, it could simplify humanoid system architecture and reduce cost, shifting the field from hardware-heavy to intelligence-driven competition. The team has not disclosed further technical details yet, and open questions include how far the capability generalizes beyond the demo and whether there are open-source or productization plans. With the humanoid market heating up, both hardware bodies and software brains are becoming entry points for founders.