July 2026 Summaries

Rescale Workflows now supports cloud-storage-based file sharing between job and workstation steps, allowing engineering teams to connect preprocessing, simulation solving, and post-processing activities within a single reusable workflow. Announced by Rachel Fu on July 31, 2026, the feature aims to eliminate manual file hand-offs and the need to configure separate storage systems while supporting deterministic multi-step simulation processes.

Rescale Workflows facilitates seamless file sharing across multi-step simulation processes by utilizing cloud storage to transfer files directly between job and workstation stages without requiring manual intervention or additional storage configurations. This integration allows engineering teams to maintain continuity across preprocessing, solving, and post-processing phases within a single, reusable workflow. The streamlined approach enhances efficiency by keeping all stages connected and reduces the complexity of managing separate storage solutions, as demonstrated in a demo showcasing the capabilities of Rescale Workflows for deterministic simulation processes.

Rescale's innovative Commitment Plans and orchestration tools provide engineering teams running high-performance computing (HPC) workloads in the cloud with enhanced cost-efficiency and flexibility, addressing the limitations of traditional cloud cost-saving mechanisms. These plans allow organizations to implement savings strategies that align with their specific computing needs, providing significant discounts without compromising on performance. Key features include automatic cost alignment, visibility into spending, and policy-driven controls, enabling administrators to manage costs effectively while maintaining operational workflow. The Rescale Savings Plan offers a simplified, dollar-per-hour commitment that applies across various coretypes and regions, ensuring savings even for varied and unpredictable workloads. Additionally, Rescale's Coretype Plans cater to predictable workloads by offering deeper discounts, while Coretype Collections simplify hardware selection through curated bundles. These offerings are underpinned by Rescale's cloud-agnostic infrastructure, which optimizes access to cutting-edge, cost-efficient cloud resources, thereby transforming compute economics from a byproduct of technical decisions into a strategic operational discipline.

Agent-accelerated computational chemistry on Rescale enhances molecular dynamics simulations by providing agentic support that aids researchers in analyzing results, proposing subsequent steps, configuring new simulations, and critiquing outcomes while maintaining scientific context. This approach integrates high-performance computing execution, scientific analysis, historical knowledge, and job setup into a cohesive workflow, ensuring researchers remain in control. A demonstration of the agents in action is available to illustrate their role in computational chemistry workflows, offering insights into the broader implications for scientific and engineering research and development on the Rescale platform.

Rescale’s Agent-Accelerated Computational Chemistry workflow uses agentic support to help researchers run and interpret molecular dynamics simulations more efficiently. The agents can analyze simulation results, recommend next steps, configure follow-up jobs, and critique outcomes while retaining scientific context across HPC execution, analysis, historical knowledge, and simulation setup. The approach is designed to provide guided support for scientific R&D workflows while keeping researchers in control, with additional demonstrations and information available through Rescale.

Rescale's updated AI Physics workflow offers an innovative approach for building surrogate models in automotive crash and structural simulations by integrating GeoTransolver support for transient finite element analysis (FEA) data. This enhancement facilitates the extraction of solver-specific metadata and provides training-ready formatting within a reusable pipeline, significantly reducing the need for manual setup and allowing engineers to achieve quicker crash insights. The initiative aims to streamline the simulation process, and resources such as tutorials and documentation are available to deepen understanding and practical application of AI Physics in this context.

Rescale’s AI Physics workflow enables engineers to build surrogate models for automotive crash and structural simulations using transient finite element analysis data. The updated crash use case integrates GeoTransolver support for transient behavior, solver-specific metadata extraction, and training-ready data formatting into a reusable pipeline, aiming to reduce manual preparation and accelerate simulation insights. Additional tutorials and resources are available through the AI Physics documentation.

Rescale's Agent Fidelity Toolkit (RAFT) offers a comprehensive evaluation framework for engineering agents, ensuring their reliable deployment. By integrating response scoring, tool-use validation, ground-truth checks, and continuous monitoring, RAFT systematically verifies the effectiveness and accuracy of agents' responses and actions. This approach addresses the quality requirements essential for agentic engineering in production environments, promoting confidence in the agents' performance.

Rescale outlines an evaluation framework for deploying engineering agents confidently through its Rescale Agent Fidelity Toolkit (RAFT). RAFT combines response scoring, tool-use validation, ground-truth checks, and continuous monitoring to assess whether agents provide relevant answers and perform correct actions. The approach is intended to establish the quality controls needed for reliable production use of agentic engineering systems.

Rescale Interlink is an open-source hybrid command-line and graphical application designed to streamline simulation data transfers and job management between desktop systems and the Rescale cloud platform. Available for Windows, macOS, and Linux and compliant with FIPS 140-3, it provides multithreaded parallel file transfers, a visual file browser, automatic downloading of completed job results, and job submission capabilities in a lightweight community-driven tool.

Rescale Interlink is an open-source tool designed to facilitate faster movement of simulation data between desktops and the cloud, combining both CLI and GUI functionalities for efficient file and job management on the Rescale platform. It is FIPS 140-3 compliant and compatible with Windows, Mac, and Linux, offering features such as multithreaded parallel transfers, a visual file browser, automatic download of completed job results, and comprehensive job submission capabilities. This lightweight, community-driven application aims to streamline data transfer processes and is further detailed in its public GitHub repository.

Rescale's Job Troubleshooting Agent streamlines the process of resolving simulation failures by quickly analyzing solver logs, identifying the issues, and providing clear explanations and recommended corrective actions, all accessible within the Rescale Assistant. This tool allows engineers to maintain their oversight by reviewing and approving suggested fixes before resubmission, thus reducing the manual workload typically associated with diagnosing failures. The service highlights Rescale's specialized agents designed to support common engineering use cases, aiming to enhance efficiency by transforming what could be hours of troubleshooting into mere minutes.

Rescale’s Job Troubleshooting Agent is an AI capability within the Rescale Assistant that helps engineers diagnose simulation failures more quickly by analyzing solver logs, identifying likely causes, explaining issues in plain language, and recommending corrective actions. Engineers retain control by reviewing and approving proposed fixes before resubmitting jobs, combining automated diagnosis with human oversight to reduce the manual effort and time typically required for troubleshooting.

Rescale’s Simulation Monitor automation is designed to identify diverging or stalled iterative solver simulations in real time, helping users avoid wasting compute resources. Running alongside jobs on the Rescale platform, it provides a live dashboard with residual plots, switchable tabular data, and automated CSV exports without requiring an interactive workstation. The automation currently supports Ansys Fluent, STAR-CCM+, and CFX, while its plugin-based architecture is intended to enable rapid support for additional simulation solvers.

Simulation Monitor automation on Rescale aids in detecting diverging and stalled simulations in real-time, preventing unnecessary computational expenditure by running alongside iterative solver jobs. It provides users with a live dashboard featuring residual plots, tabular data switching, and automated CSV exports, eliminating the need for an interactive workstation. Currently supporting Ansys Fluent, STAR-CCM+, and CFX, the system is built on a plugin-based framework that allows for quick integration of additional solvers. This automation exemplifies how Rescale's innovations can streamline routine modeling and simulation tasks.

Rescale has expanded its data connectors to streamline access to external engineering data, allowing teams to integrate and utilize simulation inputs, reference documents, and program files from existing storage systems such as AWS S3, Azure Blob Storage, and SharePoint without migrating data. This enhancement facilitates easier browsing, searching, and importing of files directly into Rescale jobs and workstations, eliminating unnecessary manual steps and improving workflow efficiency. The platform's search capabilities, including semantic search and path filtering, enable engineers to find relevant information quickly, enhancing both operational tasks and complex analysis. By structuring connected data into a searchable layer, Rescale supports agentic and AI Physics workflows, ensuring traceability and reliability in engineering processes. These improvements aim to bridge the gap between where data resides and where it is needed, reinforcing the platform's foundation for advanced simulation and AI applications.

Rescale AI Physics now offers access to the open-source DrivAerML dataset, enabling engineers to begin training surrogate models for automotive aerodynamics without first creating their own simulation data. Derived from a high-fidelity public dataset containing 500 parametrically morphed vehicle variants, the resource is intended to accelerate experimentation and adoption of AI Physics model training workflows.

Open-source AI datasets for AI Physics model training are now available on Rescale, facilitating quicker development of surrogate models. Engineers can utilize the DrivAerML dataset, a high-fidelity public dataset focused on automotive aerodynamics, which includes 500 parametrically morphed variants. This resource enables users to commence AI Physics model training without the need to generate their own simulation datasets, offering a streamlined entry point for those interested in AI Physics.

Agentic Digital Engineering, recently introduced by Rescale, aims to revolutionize AI-first product development by enhancing AI-assisted workflows, operationalizing product development, and optimizing trade-offs between speed, throughput, and cost within digital engineering. This innovation includes simulation-native agents that automate processes such as input validation and failure diagnosis, and an AI physics operating system that transforms simulation data into surrogate models, enhancing cost-efficiency and design evaluation. Rescale's collaboration with U.S. national laboratories is set to leverage agentic AI for manufacturers, and their presence at various international events highlights the integration of AI into engineering workflows without added complexity. Additionally, updates to Rescale's platform and partnerships are expanding access to high-performance computing architectures and the latest simulation software, supporting the growing demand for AI-driven engineering solutions.

GeoTransolver, now available in Rescale AI Physics through the NVIDIA PhysicsNeMo library, is a transformer-based architecture for developing high-accuracy AI surrogate models for computational fluid dynamics and finite element analysis. Using Geometry-Aware Latent Embedding attention, it captures relationships within complex three-dimensional geometries and unstructured meshes while generalizing across operating conditions and transient physical behavior. The model is intended to support predictions of aerodynamic, structural, and deformation outcomes across design variants, illustrated by a surrogate model that assesses a soccer ball’s contact and deformation after striking a crossbar to determine whether it results in a goal.

GeoTransolver is a geometry-aware transformer architecture that integrates into the Rescale AI Physics platform as part of the NVIDIA PhysicsNeMo library, enhancing the development of high-accuracy AI surrogate models for Computational Fluid Dynamics (CFD) and Finite Element Analysis (FEA). Utilizing GALE attention, it effectively models relationships across intricate 3D geometries and unstructured meshes, allowing it to adapt to varying operating conditions and comprehend transient, geometry-aware physical behaviors. This makes GeoTransolver particularly valuable for predicting aerodynamic, structural, or deformation behavior across different design variants, thereby offering significant advancements in surrogate model development within the realm of AI-driven physics applications.