Double Omniverse Isaac Sim NVIDIA GPUs on AWS Help Robots

The pace at which more intelligent robots are developed on the cloud is about to accelerate

With the arrival of NVIDIA Isaac Sim and NVIDIA L40S GPUs on Amazon Web Services, developers will be able to create and implement accelerated robotics applications in the cloud. Developed on the NVIDIA Omniverse development platform for creating and integrating OpenUSD applications, Isaac Sim is an extendable simulator for robots with artificial intelligence capabilities.

The L40S GPU is designed to power the workloads of the next generation of data centers by combining powerful AI compute with media acceleration and graphics. The L40S, which is based on the Ada Lovelace architecture, boosts engineering and robotics teams by enabling rapid real-time rendering and giving up to a 3.8x performance leap for Omniverse compared with the previous iteration.

Using Isaac Sim, the generational leap in acceleration yields 2x quicker performance than the A40 GPU across a wide range of robotic simulation workloads.

Additionally, L40S GPUs can be used for generative AI tasks, such as real-time inferencing for chat and text-to-image applications, and fine-tuning massive language models in a matter of hours.

Roboticists will be able to quickly access preconfigured virtual machines to run Isaac Sim workloads thanks to new Amazon Machine Images (AMIs) on the NVIDIA L40S in the AWS Marketplace.

The advancement of robotics in simulation is accelerating application deployment, boosting sectors like retail, food processing, manufacturing, logistics, and more.

According to ABI Research, revenue from mobile robots in warehouses is predicted to soar, more than triple from $11.6 billion in 2023 to $42.2 billion by 2030.

Robotic devices have been crucial in helping fulfillment centers satisfy the demands of online buyers while also giving workers a better work environment. Over 750,000 robots have been installed by Amazon Robotics in its facilities worldwide in an effort to enhance the working conditions for both its customers and staff who assist package fulfillment.

“They develop, test, and deploy they robots in a way that depends heavily on simulation technology,” stated Brian Basile, head of virtual systems at Amazon Robotics. They at Amazon Robotics are always expanding the scope and intricacy of they models. They will continue to push the limits of rendering, simulation, and model training with the new AWS L40S offering.

Using Isaac Sim to Hasten the Development of Robots

Large datasets may be required by robotics systems in order to operate precisely in deployed applications. Acquiring these datasets and putting them through real-world testing is an expensive, time-consuming, and unrealistic process.

AI-based robotic applications are trained and tested using robotics simulation. Simulations are making virtual advancements possible like never before with synthetic data. Prior to operation, simulations can aid in the verification, validation, and optimization of robot designs, systems, and algorithms. In order to maximize efficiency and lower the cost of manufacturing change orders, it can also be utilized to optimize facility designs prior to the commencement of construction or remodeling.

Teams may work together more successfully using Isaac Sim’s cloud access and access to the newest robotics simulation tools and features. Machine learning engineers may create production-ready synthetic datasets for training resilient deep learning perception models in Isaac Sim by utilizing the Omniverse Replicator synthetic data generation engine.

Adoption of Isaac Sim by Customers on Amazon

Among the AWS early adopters of the Isaac Sim platform are Theory Studios, Amazon Robotics, and Soft Robotics.

In order to automate, optimize, and plan its autonomous warehouses in virtual environments before deploying them in the real world, Amazon Robotics has started utilizing Omniverse to create digital twins.

Amazon Robotics will improve its Proteus autonomous mobile robot and expedite its development by using Isaac Sim for sensor emulation. This will help the online retail giant handle fulfillment more effectively.