Log in Sign up
Coding Assistants A 7.89 GB Two-Bit Quant of Qwen3.8-27B Beats the 54 GB Original at Tool Calling AI Agents AWS Open-Sources a Physical AI Toolchain That Wires SageMaker to Nvidia Isaac Sim, GR00T and Cosmos News The Best Video Model Scores Just 57.76 Out of 100 on a New Physics Benchmark Built From 40 Measured Experiments Business Biren Technology Returns to the Market a Third Time in Ten Months With a $515 Million Share Sale News Singapore's MAS Turns AI Guidance Into Binding Rules for Banks, Phased In Through 2028 Business Hone Raises $60 Million Seed at a Reported $285 Million to Sell Agents That Own an Outcome Coding Assistants AI Coding Agents Add 30% More Code but Not More Software, Harvard Study Finds Security One Prompt, an Entire Region of Agents: Inside the Patched AgentCorruption Flaw in AWS Bedrock AgentCore Coding Assistants A 7.89 GB Two-Bit Quant of Qwen3.8-27B Beats the 54 GB Original at Tool Calling AI Agents AWS Open-Sources a Physical AI Toolchain That Wires SageMaker to Nvidia Isaac Sim, GR00T and Cosmos News The Best Video Model Scores Just 57.76 Out of 100 on a New Physics Benchmark Built From 40 Measured Experiments Business Biren Technology Returns to the Market a Third Time in Ten Months With a $515 Million Share Sale News Singapore's MAS Turns AI Guidance Into Binding Rules for Banks, Phased In Through 2028 Business Hone Raises $60 Million Seed at a Reported $285 Million to Sell Agents That Own an Outcome Coding Assistants AI Coding Agents Add 30% More Code but Not More Software, Harvard Study Finds Security One Prompt, an Entire Region of Agents: Inside the Patched AgentCorruption Flaw in AWS Bedrock AgentCore

AWS Open-Sources a Physical AI Toolchain That Wires SageMaker to Nvidia Isaac Sim, GR00T and Cosmos

AI Agents · By: AIReport Admin · 1 hour ago · 👁 6
AWS Open-Sources a Physical AI Toolchain That Wires SageMaker to Nvidia Isaac Sim, GR00T and Cosmos

Amazon Web Services launched an open-source Physical AI Toolchain on October 8 that combines SageMaker, EC2 GPU instances, S3, IoT Greengrass and Bedrock AgentCore with Nvidia Isaac Sim, Isaac Lab, Isaac GR00T and Cosmos. AWS says the architecture draws on more than 1 million robots it has deployed across its own network.

Amazon Web Services has released an open-source Physical AI Toolchain that packages the plumbing between cloud infrastructure and robots into reference architectures, infrastructure-as-code templates and deployment automation. The announcement, made on October 8, is aimed at the least glamorous part of robotics: the engineering work that stands between a trained model and a machine that can actually run it.

"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," said Uwem Ukpong, vice president of AWS Industries. "We want to flip that." The toolchain is organized around five stages: synthetic data generation, model training, simulation and validation, edge deployment, and continuous improvement using operational data from machines already in the field.

On the AWS side, the reference architecture identifies Amazon SageMaker for model training, Amazon EC2 GPU instances for simulation, Amazon S3 for data storage, AWS IoT Greengrass for edge deployment and Amazon Bedrock AgentCore among the orchestration capabilities. On the Nvidia side it incorporates Isaac Sim for simulation, Isaac Lab for reinforcement learning, Isaac GR00T for humanoid robot training and Cosmos for synthetic data generation, with Nvidia OSMO handling workflow orchestration.

Developers are not locked into the full stack. The toolchain accepts custom robot descriptions and teleoperation data in common formats, including PyTorch, Hugging Face, Gymnasium, ROS 2, the LeRobot data format, URDF and ONNX, and each component can be used on its own. One shipped example pairs Isaac GR00T with 27 episodes of UR3 robot pick-and-place teleoperation data to demonstrate the workflow end to end.

AWS is explicit that this is a distillation of its own operations: the company says it has deployed more than 1 million robots across its network, where they move millions of packages a day alongside hundreds of thousands of employees. It also cites an AWS Startups and Strand Partners report finding that one in seven startups is building physical AI and that 72 percent of those builders consider cloud computing essential to their systems.

Early users named in the announcement include NEURA Robotics, which is developing cognitive humanoid robots; RLWRLD, which is building an 8.1-billion-parameter foundation model for dexterous manipulation; and Config, which has assembled a pipeline of more than 200,000 hours of robot-action data and uses generative AI to synthesize additional training scenarios. "In Physical AI, speed is everything," said NEURA founder and CEO David Reger. Nvidia's Amit Goel, head of robotics developer ecosystem and edge AI product, framed the same problem as the need to integrate three computing platforms: training, simulation and deployment.

AWS targets industrial automation, warehousing and logistics, energy, healthcare, mining, agriculture, aerospace and defense, and says manufacturers can use the toolchain to reach production in weeks rather than the years a bespoke pipeline can take. That timeline is the company's stated expectation, not an independently verified result. The substantive claim underneath is narrower and more credible: most robotics teams lose months to integration work that has now been written down once and open-sourced under the Apache 2.0 license.

Comments (0)

Log in to join the discussion

Log In

No comments yet