Humanoid Robots Must Rely on Local Compute
2C_ornot2C · x · 2026-07-17
This in-depth article explores why humanoid robots must carry their "brains" onboard rather than relying on cloud-based inference.
The core argument is that robot-environment interactions happen in milliseconds. Cloud round-trips of 100–500ms are simply too slow, leading to missed grasps, collisions, and skipped actions. Therefore, critical modules like perception, VLA, and control loops must run locally on onboard chips.
The article notes that many current humanoid robots are designed around chips like the Nvidia Jetson Thor: roughly 2000 TFLOPS, 128GB of memory, and 130W power consumption, delivering about 5x the performance of the previous generation. However, 130W of compute power isn't free—it directly competes for power with the battery and actuators. Every extra watt dedicated to compute means one less watt of battery life.
More from Embodied
- Tesla expands Robotaxi rides to seven areas, including new Orlando and Tampa zones — elonmusk · 2026-07-22
- Hands-on robotics workshop on Saturday may be the last in-person session before August — StewartalsopIII · 2026-07-22
- NVIDIA pitches World Foundation Models as a way to scale physical AI data generation — MonaJalal_ · 2026-07-22
- RoboMME Podcast Preview: Benchmarking Memory for Robotic Policies — chris_j_paxton · 2026-07-21
- Gritt says an 8-person crew now installs 3,000 to 4,000 solar panels a day — HaktanSuren · 2026-07-21
- A helium-powered flying robot whale aims to be a quiet companion pet — chris_j_paxton · 2026-07-21