Humanoids run on milliseconds, not gigabytes.
Robotics is one of many applications where the substrate's properties matter — bounded latency, tiny footprint, online correction. The same plug-points work in healthcare, edge inference, conversational AI, and online recommendation.
Latency budgets are absolute
Vision, planning, balance, locomotion, manipulation, speech and intent must all close their control loops in tens of milliseconds. Cloud round-trips and giant transformer stacks simply cannot fit.
Memory and power are scarce
Humanoid platforms run on battery. Every gigabyte of RAM and every watt of GPU draw is taken from balance, vision and runtime. You need dozens of skills, not one giant model.
Reliability over hype
On a robot, accuracy without bounded latency is worthless. Operators need shapes whose worst-case behavior is known, repeatable, and small enough to deploy at the edge.