June 2026 Summaries

Rescale has introduced a feature that allows engineering data stored in AWS S3 and Azure Blob Storage to be directly connected to the Rescale Assistant via vectorized search, enabling engineers to interact with and query their data in one centralized location. This innovation allows users to ask natural language questions and obtain rapid, contextually relevant answers based on their organization's specific data, reducing reliance on general AI knowledge and eliminating the need for manual file searches.

Rescale's Report Generator Agent streamlines the process of converting raw simulation study results into structured, presentation-ready reports in significantly less time, transforming what typically takes hours into just minutes. By aligning the generated reports with a customer's specific reporting templates, complete with findings, charts, and recommendations, it eliminates the need for engineering teams to manually prepare slides or summarize data. This tool is designed to produce outputs that are immediately suitable for stakeholders, enhancing efficiency and ensuring consistency in reporting.

Rescale's Budget Control Agent offers engineering and IT teams a proactive solution for managing compute budgets in real-time by monitoring expenses, identifying potential budget risks before they escalate, and providing guidance on corrective measures using natural language through the Rescale Assistant. This tool shifts the approach from reactive to proactive, ensuring that stakeholders remain informed and have control over their spending, thus enhancing budget management effectiveness. Additionally, Rescale agents are designed to aid in advanced modeling and simulation tasks, further supporting efficient resource utilization.

The Rescale AI Physics Local Inference App is a versatile desktop application available for Windows, Mac, and Linux, enabling engineers to utilize AI models trained on the Rescale platform directly on their local hardware without needing an active cloud session. This application is particularly advantageous during design review cycles, allowing engineers and designers to interactively query surrogate models without the delay of cloud-based jobs. It operates on CPU by default but can leverage local GPU resources automatically if available, eliminating the need for additional libraries or frameworks. The app also facilitates local inference workflows for teams working in on-premises environments or those with network restrictions, enhancing flexibility and efficiency in digital engineering processes.

Cloud High Performance Computing (HPC) is transforming digital twin and multi-physics simulation performance, driven by the increasing complexity and volume of simulation iterations that incorporate AI. Rescale, in collaboration with Amazon Web Services (AWS), provides engineers and scientists with the necessary compute infrastructure to support these advanced workloads, now offering the Amazon EC2 Hpc8a instance powered by 5th Gen AMD EPYC processors on its platform. The Hpc8a instance showcases significant speed improvements over its predecessor, Hpc7a, with an average 52% speedup in Computational Fluid Dynamics (CFD) applications and a 38% improvement in Finite Element Analysis (FEA), due to architectural enhancements in memory bandwidth and network latency. This new instance is designed to seamlessly enable engineers to run more high-fidelity models and iterations, facilitating innovation across industries such as automotive, aerospace, semiconductor design, and life sciences without the need for complex reconfigurations. Rescale's platform is built to ensure easy transition and immediate access to these performance gains, underscoring its commitment to advancing AI-native engineering and simulation-driven industries.

Amazon EC2 Hpc8a instances, now generally available on the Rescale platform, are designed to meet the growing demands of engineering teams that conduct high-fidelity simulations with AI integration. These instances, powered by 5th Gen AMD EPYC™ processors, offer significant performance improvements over their predecessors, notably a 52% average speedup in Computational Fluid Dynamics (CFD) applications and a 38% speedup in Finite Element Analysis (FEA) workloads. The Hpc8a's enhancements in memory bandwidth and network latency facilitate near-linear scaling and reduced bottlenecks, enabling more efficient data movement and problem-solving in complex, simulation-driven industries such as automotive, aerospace, and life sciences. Rescale's collaboration with AWS ensures that the latest HPC innovations are accessible to its customers, supporting seamless transitions to new hardware without the need for reconfiguration, thereby driving R&D forward and maximizing innovation.

At the 2026 NAFEMS Americas conference, the focus on AI for digital engineering highlighted its transition from exploratory discussions to practical deployment, with a significant portion of the program dedicated to AI's implementation in engineering workflows. The conference underscored the growing importance of agentic engineering, emphasizing the need for responsible deployment of autonomous agents that assist rather than replace human decision-making. Key discussions revolved around the challenges of integrating AI into existing workflows, particularly in terms of data accessibility and governance, with practitioners favoring narrow use cases such as error detection and job orchestration. Rescale played a prominent role, presenting on agentic workflows and AI Physics infrastructure, showcasing AI-first engineering applications and unveiling their AI Physics OS for developing and governing surrogate models. The event also highlighted a widening gap between teams with robust cloud-native infrastructures and those still in early experimental stages, with data readiness and trust at the workflow level identified as critical factors for successful AI adoption in engineering.