Toolchain launch: Amazon Web Services introduced an open-source Physical AI Toolchain on AWS combining its infrastructure with NVIDIA robotics software.
- Lifecycle coverage: The toolchain supports development, training, simulation, and deployment, with reference architectures, infrastructure-as-code resources, and deployment automation.
- Stack roles: NVIDIA Isaac Sim and Isaac Lab cover simulation and reinforcement learning, while Amazon SageMaker, Amazon EC2 GPU instances, Amazon S3, and AWS IoT Greengrass cover training, simulation, storage, and edge deployment.
- Manufacturing focus: Target sectors include automation, warehousing, logistics, energy, healthcare, mining, agriculture, aerospace, and defense, with an intent to help manufacturers launch capabilities in weeks.
Amazon Web Services has introduced an open-source toolchain for physical AI development, combining Amazon Web Services infrastructure with NVIDIA robotics software to support development, training, simulation, and deployment of physical AI systems.
The Physical AI Toolchain on AWS is built around reference architectures, infrastructure-as-code resources, and deployment automation. Developers can supply their own robot descriptions, teleoperation data, and task definitions rather than assembling every layer themselves.
“We built the Physical AI Toolchain on AWS because customers told us that too much of their engineering effort was going to infrastructure instead of innovation. We want to flip that,” said Uwem Ukpong, vice president of AWS Industries.
From simulation to edge deployment
The workflow assigns specific roles across the stack: NVIDIA Isaac Sim is used for simulation and Isaac Lab for reinforcement learning. Amazon SageMaker is used for model training, and Amazon EC2 GPU instances for simulation. Amazon S3 is used for data storage, and AWS IoT Greengrass for edge deployment.
The toolchain also incorporates technologies including PyTorch, Hugging Face, Gymnasium, Robot Operating System 2 (ROS 2), Unified Robot Description Format (URDF), and Open Neural Network Exchange (ONNX). Amazon Web Services provides technical documentation and implementation materials for the toolchain.
Manufacturing intent
The targeted industrial sectors include automation, warehousing, logistics, energy, healthcare, mining, agriculture, aerospace, and defense. The toolchain is designed to help manufacturers launch physical AI capabilities in weeks, and is intended to accelerate the cycle of fine-tuning models, deploying them, and scaling to large fleets.
That emphasis on cycle time is echoed by users in the ecosystem. NEURA Robotics is developing cognitive humanoid robots using the toolchain, while RLWRLD is working on foundation models for dexterous manipulation.
“In Physical AI, speed is everything: how fast you can fine-tune models, deploy them into the real world and scale from individual systems to large fleets,” said David Reger, founder and CEO of NEURA Robotics.
Amit Goel, head of NVIDIA’s Robotics Developer Ecosystem and Edge AI Product, framed the requirement as an integration problem across compute environments.
“Building physical AI requires a seamless integration of three computing platforms—training, simulation, and deployment.”
Is this your company?
This article features your business. Claim it to add your logo, contact details, and a link to your website — or upgrade to reach more buyers.
Did you know 80% of Press Releases trigger AI content warnings? Reach out and the M4S team can assist.