Rescale expands access to NVIDIA GPU simulation and AI in cloud

Expanded collaboration provides engineers and scientists with scalable computing resources.

Rescale announced the expansion of its global network of cloud service providers to include NVIDIA DGX Cloud via the NVIDIA Cloud Partner network, enabling customers to leverage the leading-edge architectures from NVIDIA in more locations with greater access to capacity.

This collaboration expands Rescale’s global network of cloud service providers to deliver unmatched access to hundreds of cutting-edge compute services and the latest architectures. By enhancing its integration with NVIDIA DGX Cloud and launching the new Rescale CAE Hub powered by NVIDIA, Rescale delivers greater availability of GPU-accelerated simulation and AI workloads to independent software vendors’ customers. Rescale customers will benefit from enhanced performance, scalability, and cost savings through better access to the latest cutting-edge GPU infrastructure, such as the NVIDIA GB200 NVL72 platform, which is powered by NVIDIA Blackwell architecture and high-speed NVIDIA NVLink. Through this collaboration, engineers and scientists can dramatically accelerate their simulation workflows — for example, leading high-performance CFD codes deliver up to 50x faster results when run on GPU Blackwell versus traditional architectures.

Rescale’s open platform offers the control and data planes necessary for workload orchestration across multi-cloud and multi-software environments, laying the foundations for building AI Physics workflows. This is why leading companies like Boom Supersonic, the company building the world’s fastest airliner, are turning to Rescale and NVIDIA to explore new possibilities and achieve faster, more efficient innovation.


Working with Rescale and NVIDIA allows Boom to explore 4x more design possibilities, achieve 10x faster aerodynamic simulations, and drive 30x greater cost efficiency, accelerating innovation like never before.

By increasing access to cost-efficient, high performance computing, this collaboration expands Rescale’s capabilities and enhances its ability to support customers in organizing data, developing and deploying AI Physics methods for modeling & simulation. A critical enabler of this effort is infrastructure that securely and automatically captures and pre-processes simulation data and its metadata at the point of compute to generate high-quality training datasets. By building data management into the HPC workload orchestration layer, Rescale’s platform converts raw, unstructured simulation data into AI-ready training sets, accelerating innovation and deployment.

Rescale, powered by NVIDIA technologies, is driving high performance computing innovation worldwide, providing customers with tailored support and solutions to tackle their most ambitious challenges. A prime example, highlighted at last year’s GTC 2024, is GM Motorsports, which leveraged Rescale’s AI Physics platform and NVIDIA accelerated computing to achieve simulations up to 1,000 times faster than traditional methods, with over 98% accuracy — demonstrating the real-world potential of AI-driven engineering.

For more information, visit rescale.com.