Rescale, U.S. National Labs Form Initiative

Rescale announces a proposed collaboration with three U.S. Department of Energy national laboratories: Lawrence Berkeley National Laboratory (LBNL), Oak Ridge National Laboratory (ORNL), and Lawrence Livermore National Laboratory (LLNL).

The collaboration builds on existing work with the ORNL Manufacturing Demonstration Facility to bring simulation codes developed at the national labs to American industry at scale.

Operating as the Agentic HPC Pipeline Initiative (AHPI), the partnership would use launched agentic digital engineering capabilities on the Rescale platform to accelerate U.S. manufacturers’ deployment of DOE-developed codes in production engineering environments.

By hosting DOE codes on the Rescale digital engineering platform and pairing them with agentic AI that guides users through all steps, engineers in U.S. industry can translate intent into fully autonomous simulation workflows. Tasks that demand deep specialist knowledge, such as mesh configuration and hardware selection, are handled autonomously by the simulation-native agents.

“America’s national laboratories have spent decades building the most powerful engineering simulation software in the world, like our AMReX framework, but the depth of domain expertise required to configure and validate these tools has kept them out of reach for many manufacturers,” says Peter Nugent, division deputy for Science in the Applied Math & Computational Research Division at LBNL and principal investigator of the AHPI project. “We are replacing the manual burden of solver selection, checkpointing, and hardware configuration with intelligent automation.”

Rescale’s in-platform agents are orchestrating manufacturing codes developed at ORNL’s Manufacturing Demonstration Facility.

“At ORNL, we are focused on pushing the boundaries of what is possible in physical production,” says Ramanan Sankaran, ORNL lead. “By reducing the expertise bottleneck in job configuration and allowing engineers to train surrogate models on their simulation data using in-platform agents, we are enabling U.S. manufacturers to innovate at speeds that were previously impossible.”

The proposed collaboration with LLNL builds on the established HPC for Energy Innovation (HPC4EI) program, which connects U.S. companies with national laboratory expertise and resources in high-performance computing, modeling and simulation.

“Through HPC4EI, LLNL has a long and successful history of bridging the gap between national lab supercomputing and real-world industrial applications,” says Aaron Fisher, LLNL lead. “We believe the use of agentic AI will lead to a step change in our ability to bring these capabilities to American industry.”

Together, the three laboratories bring depth across fluid dynamics, phase and chemical transformations, additive manufacturing, materials science, and energy systems, covering advanced manufacturing applications. The Rescale platform is intended to provide secure, cloud-native access to these codes through an intuitive, agent-augmented interface that requires no specialized HPC expertise to operate.

“In the United States, both industry and the national labs have made extraordinary investments in computing infrastructure and simulation software,” says Joris Poort, CEO of Rescale. “Rescale’s digital engineering platform serves as the bridge that brings both investments together in one place and puts them to work accelerating innovation for American industry. By pairing the world’s best simulation codes with AI agents that automate the complex workflows around them, we are giving American manufacturers a compounding advantage to accelerate progress, which is the ability to explore design spaces, validate materials, and optimize manufacturing processes at speeds and scales that were previously out of reach.”

The Rescale platform, including the capabilities powering the AHPI consortium, is available now. 

Sources: Press materials received from the company and additional information gleaned from the company’s website.