Supply Chain Optimization
Agility Robotics needed a reliable way to train whole-body control for humanoid robots that could work in human-built warehouses and factories without costly facility redesigns or risky trial-and-error learning on the production floor.
Reported impact
Agility Robotics cut controller iteration cycles from weeks to days, trained Digit across billions of simulated interactions, and transferred whole-body control policies from digital twins into live warehouse and manufacturing operations with better stability, higher reliability, and faster deployment readiness.
Agility Robotics built a simulation-first training stack for Digit by combining NVIDIA Isaac Sim, Isaac Lab, NVIDIA Omniverse, OpenUSD, and GPU-accelerated reinforcement learning. Engineers model Digit's chassis and contact surfaces in Isaac Sim, then ingest customer CAD and BIM data into shared OpenUSD scenes that capture aisle widths, floor friction, lighting, and shove forces. These digital-twin scenarios flow into Isaac Lab, where millions of parallel reinforcement-learning episodes train a whole-body controller on NVIDIA GPUs and keep perception tightly coupled with motion planning. Agility then validates the same policy in a containerized MuJoCo pipeline to expose simulator-specific corner cases before deployment to real robot hardware. This end-to-end workflow lets Digit learn balance recovery, obstacle avoidance, and material-handling behavior in simulation so trained controllers transfer into live warehouse and manufacturing operations without lengthy manual retuning.
Robot geometry plus customer CAD and BIM data are modeled in Isaac Sim and OpenUSD scenes, simulation episodes flow into Isaac Lab for GPU-accelerated reinforcement learning, validated policies are cross-checked in MuJoCo, and the resulting controllers are deployed onto Digit robots in warehouse and manufacturing environments.
Digital twin simulations flow into Isaac Lab, where millions of parallel reinforcement-learning episodes refine Digit's whole-body controller.
Engineers begin by modeling the robot's chassis in Isaac Sim, using OpenUSD to preserve every joint, mass, and contact surface.
Isaac Sim running on NVIDIA GPUs lets us simulate years of real-world learning for Digit in just hours.
Iteration cycles for developing and testing new controllers are reduced from weeks to days through multi-GPU physics-based simulation before real-world deployment.
NVIDIA GPUs push training into the billions of interactions that true dexterity demands.
Review how this reference was scored and trace each normalized claim back to its documented source.
Scoring methodology
Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.
Traceability
Field coverage shows which facts each source supports. Dates are shown only when documented by the source.
The publication says Agility is headquartered in Salem, Oregon and is currently employing 371 people.
Related
NVIDIA Sample Sep 22, 2026 Sluicebox Sluicebox needed to turn fragmented bills of materials, supplier records, PDFs, CAD files, and emissions inventories into timely product carbon assessments, but manual life-cycle-analysis work was slow, expensive, and… 51-80% lower inference costs Lucy · NVIDIA Nemotron 3 Ultra · carbon graph 
In-house Sample Aug 13, 2026 Lockheed Martin Static flight-hour tables and usage metrics ignored real-time consumption, causing chronic under-resourcing, limited stock effectiveness, and a reactive sustainment supply chain that could not keep pace with modern… predict component failure up to 72 hours before it occurs LogC2 dashboard · deep-learning models · digital data fabric 
EY Sample Archived · May 11, 2026 Daikin Daikin's acquisition-driven growth left it with diverse workflows, customized systems and fragmented data across order management, supply chain, fulfillment and finance, making it hard to deliver a consistent dealer… 10% improved counter efficiency AI-assisted implementation · SAP Business Technology Platform (BTP) · SAP S/4HANA 