Figure AI Runs 50 Hours Nonstop, China IDs Its Robots, and Korea Bets Wearable Data Wins Humanoids

News

China Launches Humanoid Robot ID Card System for Full Lifecycle Tracking

Hubei province is rolling out 29-character unique IDs for humanoid robots, tracking everything from joint wear and battery status to maintenance history and liability records in real time. It is a smart infrastructure play: as China ships 84% of global humanoid units, a common traceability layer makes the secondary market, fleet management, and fault accountability actually workable at scale.

Funding

WIRobotics Closes $68M Series B to Turn Wearable Gait Data Into Humanoid Advantage

South Korean startup WIRobotics is betting that 3,000 deployed wearable walking-assist robots give it a proprietary human motion dataset that competitors cannot easily replicate for its humanoid platform ALLEX. The approach is sharp: instead of starting from scratch on manipulation data, they are converting years of real-world gait and force data into the control foundation for a full humanoid.

WIRobotics Series B Details: NVIDIA Physical AI Fellowship, AWS Partnership, Automotive PoC

The company is also in proof-of-concept discussions with a global automotive manufacturer and was selected for NVIDIA's Physical AI Fellowship, which puts it inside the supply chain conversations that matter most right now. Q1 2026 revenue already beat all of 2024, which suggests the wearable side is self-funding the humanoid R&D bet.

Tools

FANUC and NVIDIA Build Digital Twins Where Simulation and Physical Robots Behave Identically

The updated FANUC-NVIDIA integration connects ROBOGUIDE with Isaac Sim so that virtual robots run the exact same control algorithms as physical machines, keeping trajectories and cycle times matched between sim and deployment. Closing the sim-to-real gap at the industrial robot level is a big deal because it collapses commissioning time and opens reinforcement learning to factories that previously could not absorb the trial-and-error cost.

Analysis

The Figure AI Demo Was Compelling But One Expert Says It Is Not Deployment-Ready Yet

The honest read on the viral Figure AI stream: impressive uptime in a tightly controlled loop with identical packages on a repeating conveyor, but real logistics environments throw SKU variability, irregular surfaces, and mixed-weight loads at robots every shift. Worth watching the skeptics here because the gap between a 50-hour loop test and a production SLA is still wide.