2025 in Review: A Year of Platform Innovation that Drove Customer Results
As we close out 2025, we reflect on a transformational year for Rescale’s digital engineering platform. This year marked a big expansion from a cloud high performance computing (HPC) infrastructure provider to an AI-first platform for engineering and scientific innovation. We delivered comprehensive capabilities across compute elasticity, data intelligence, AI-powered productivity, and physics-based machine learning that fundamentally change how engineering teams design, simulate, and innovate.
This recap focuses on Rescale’s big investment areas that drove customer innovation in 2025, organized around four key capabilities (links to jump ahead):
Beyond these categories, we enhanced performance and ease-of-use improvements across the entire Rescale platform, continuously striving to deliver a more intuitive and snappy product experience. Looking back on all the product innovations of the year, we are most excited to have unified many of the previously disconnected parts of the modeling and simulation experience and enhanced them with integrated AI. This is the result of rapid iteration and commitment to building a foundation to support the modern capabilities our customers need to meet their product development goals.
Let’s dive into the highlights.
Delivering Turnkey, Resilient HPCaaS
Contents
Modern engineering teams need computing infrastructure that scales instantly, optimizes costs automatically, and eliminates operational overhead. In 2025, we enhanced the Rescale platform to deliver truly elastic HPC with intelligent resource management and transparent cost controls.
Expanded Access to Leading Computing Architectures
Customers now have more choice, better economics, and faster access to cutting-edge GPU infrastructure for both simulation and AI workloads. Through our expanded partnership with NVIDIA’s Cloud Partner Network, we’ve broadened our computing ecosystem to deliver GPU-accelerated simulation for computer-aided engineering (CAE) and AI in more global locations with greater capacity access. This means enhanced performance, better scalability, and cost savings through the latest GPU architectures for traditional simulation and AI workloads.
Intelligent Cost Management and Transparency
Managing cloud costs effectively remained a top priority for IT administrators in 2025, especially as teams scaled HPC usage across departments and projects. We introduced controls for multi-workspace budgets to manage consumption, allowing organizations to manage budgets for teams spanning multiple workspaces as a single entity rather than tracking separate budgets for each workspace. Combined with precise usage budget management and enhancements for granular billing & cost management that included API-driven usage reporting and detailed consumption reports, administrators gained unprecedented visibility and control over spending patterns. This means IT teams can proactively prevent budget overruns, allocate costs accurately across departments, and optimize resource allocation based on actual usage data.
Balancing cost control with engineering productivity is an ongoing challenge and teams need fast access to resources, but unmanaged hardware and software can lead to budget overruns. Refreshed Hardware Policies gives admins the ability to define and enforce compute resource constraints, ensuring teams operate within approved hardware configurations while maintaining the flexibility engineers need. These policy-driven controls automate governance without creating friction, enabling IT to prevent costly misconfigurations while engineers have instant access to the right resources.
Streamlined Software Deployment
Deploying custom and proprietary applications in cloud environments has traditionally been complex and risky, requiring careful testing, license management, and coordination to avoid disrupting engineering workflows. Rescale Software Publisher, now generally available for all customers, enables organizations to deploy custom and proprietary applications across the Rescale platform with confidence. Now with the ability to preserve sandboxes, admins can ensure testing environments remain intact during the publication process, reducing deployment risk. License hosting enhancements for streamlined administration introduced automated license expiration notifications, custom maintenance windows, and vendor daemon update controls, giving administrators fine-grained control over their licensing infrastructure while minimizing service disruptions.
Modern Platform Experience
As Rescale’s capabilities have expanded into data intelligence and AI, we redesigned our interface to make these advanced tools more accessible alongside core HPC workflows. Our new platform experience, launched in 2025, delivers streamlined navigation that puts compute, data analysis, and AI capabilities in a single, cohesive experience. This modernized interface makes it easier to move between HPC jobs, data insights, and AI workflows without switching contexts. Thanks to ongoing feedback from both engineering and IT power users, we’ve made several improvements to every-day tasks and integrated new features for data and AI workflows. The new interface is now available to opt-in for all customers – we’d love to hear your feedback as we continue refining it.
R&D Data Foundation
Engineering teams spend up to 30% of their time managing data rather than innovating. Rescale addressed this challenge in 2025 by transforming how simulation data is captured, organized, connected, and analyzed throughout the product development lifecycle.
Rescale Data Intelligence Launch
Simulation data has traditionally been siloed, unstructured, and difficult to analyze at scale, making it nearly impossible for teams to extract insights across projects or leverage AI for decision-making. Rescale Data Intelligence, launched in 2025, solves this by unifying simulation data into structured, actionable insights and accelerating decisions through AI-powered analysis. Automated workflows integrate seamlessly with existing engineering processes, while Rescale’s unique automated metadata extraction from over 20 of the most common engineering solvers eliminates manual tagging and ensures consistent data organization. This means engineering teams can now capture and compound the value of every simulation moving forward, building an AI-ready data foundation.
Connecting Engineering Ecosystems
Engineering data rarely lives in just one place – it’s scattered across cloud storage, product lifecycle management (PLM) systems, and departmental silos, making it difficult to connect simulation results with broader product context. System of record Connectors for AWS S3, Azure Blob Storage, and PLM systems enable seamless integration with existing tools, so teams can enrich simulations with broader product context. In addition to being a standard source for basic job file inputs, Rescale’s investment in connectors creates a unified digital thread that spans the entire engineering lifecycle, from concept through production. We are actively testing out new connectors for an expanded ecosystem of enterprise tools to extend our Rescale Data Intelligence capabilities to more data sources and uncover even more value from historical data.
Intelligent Storage and Data Management
As simulation data volumes grow, managing storage costs and visibility becomes increasingly challenging for IT teams. In 2025, data storage became smarter and more cost-effective through multiple enhancements. Investments in foundational storage management and reporting introduced near real-time storage usage updates that refresh hourly instead of daily, improved CSV downloads with enhanced filtering, and interactive usage tables for better visibility. Customers can now add files directly to Cloud File System (CFS), streamlining workflows by enabling direct uploads for shared storage and collaboration. Additionally, automated file deletion upon archival in CFS ensures smarter, policy-driven data lifecycle management that helps teams stay within storage limits automatically.
AI Engineering
Engineering teams face mounting pressure to deliver more innovations faster, but traditional workflows haven’t kept pace. In 2025, Rescale delivered capabilities that accelerate everyday engineering tasks through intelligent automation, contextual insights, and rapid results generation, multiplying productivity and accelerating project timelines.
AI-Powered Insights
Finding relevant information in thousands of simulation results typically requires a manual searching through files and folders – a time-consuming process that slows decision-making. Rescale Assistant enables conversational data exploration and analysis, allowing engineers to query their simulation data using natural language. The assistant generates intelligent summaries of job results, surfaces patterns and anomalies, and makes finding relevant past work as simple as asking a question. AI-driven job summaries automatically highlight key findings, detect anomalies, and diagnose failures, transforming technical results into actionable insights.
Rescale Assistant will soon enable administrators to query billing data, analyze platform usage, and visualize spending trends through natural language interactions.
Foundations for Simulation Agents
Engineering teams spend significant time on repetitive tasks – setting up simulations, validating inputs, configuring design of experiments (DOE), and generating reports – work that follows predictable patterns but still requires manual effort. In 2025, we established the critical foundation for autonomous engineering through the Rescale Model Context Protocol (MCP) server and platform integrations. This infrastructure enables simulation agents to perform complex, multi-step workflows that follow organizational methods, capture project context, and embed domain expertise directly into processes.
These agents represent the next evolution of engineering automation: systems that don’t just execute predefined scripts, but learn from organizational patterns, adapt to changing requirements, and proactively orchestrate actions across tools and teams. Early implementations demonstrated agents handling routine simulation setup, input validation, DOE configuration, and automating tedious reporting tasks, keeping engineers in control while eliminating repetitive manual work.
The Rescale MCP server enables customers to connect their preferred AI tools (such as Claude, ChatGPT, or other language models) directly to Rescale’s simulation environment. Engineers can now manage simulations, query data, and automate workflows using the AI interfaces they already use daily rather than switching between multiple applications.
Automated Workflow Orchestration
Complex engineering workflows often require manually coordinating multiple tools and steps, making it difficult to standardize processes or scale best practices across teams. Engineering teams can now visually design and automate complex, multi-step tasks through no-code interfaces that capture expert methods, standardize best practices into reusable templates, and scale institutional knowledge across programs. These capabilities seamlessly integrate with the broader platform ecosystem, allowing users to invoke jobs, workstations, processing scripts, and AI agents from unified canvases. The ability to build training data pipelines for surrogate models directly within automated workflows creates a natural bridge between traditional simulation and AI-accelerated methods.
AI Physics Methods
Traditional CAE methods can require hundreds or thousands of simulations to explore a design space – a process that’s time-consuming and computationally expensive, while AI-powered approaches have historically required specialized machine learning expertise. In 2025, Rescale AI Physics made AI-powered physics simulation accessible to everyday CAE engineers, enabling them to build production-ready AI models without deep AI expertise, dramatically accelerating design space exploration.
Complete AI Operating System for Engineering
Building AI models for engineering applications traditionally requires stitching together disparate tools and manually managing complex workflows across data preparation, training, deployment, and iteration. Rescale AI Physics delivers the only end-to-end operating system for modeling and simulation, handling the full range of necessary steps including data generation, structuring, dataset creation, model training, evaluation, deployment, inference, and fine-tuning. This integrated approach transforms simulation data into powerful AI models through automated workflows that maintain traceability and version control at enterprise scale.
Engineers can now explore large design spaces rapidly with real-time AI inference models that deliver results in seconds instead of days, achieving accuracy matching traditional simulation methods while enabling rapid iteration and optimization.
Proven Customer Impact at Scale
At NVIDIA GTC 2025, Rescale showcased breakthrough AI Physics implementations with industry leaders, demonstrating how advanced simulation technologies combined with GPU acceleration and AI-powered methods deliver transformational results across aerospace and energy sectors.
Boom Supersonic leveraged Rescale’s AI Physics capabilities to accelerate the development of the Overture supersonic airliner, transforming their aerodynamics development process through AI-powered design exploration. Read the full details of the collaboration at GTC to learn how Boom Supersonic achieved:
- 4x more design iterations
- 10x GPU speedup across CFD simulations
- 400% more design possibilities explored
SLB (fka Schlumberger) demonstrated how GPU-accelerated simulation and AI-enhanced methods revolutionize energy sector workflows, particularly for computationally intensive coupled CFD-DEM simulations critical to oil and gas operations. Read the full details of the joint presentation, covering how SLB achieved:
- 10x faster simulation results with 30x cost reduction
- 3,600x speedup for field operations
- Order-of-magnitude improvements in simulation efficiency
These customer deployments validate that AI Physics delivers production-ready results across diverse physics domains while maintaining the accuracy engineers require for critical design decisions.
Expanded Surrogate Model Support
Building surrogate models traditionally requires choosing between off-the-shelf solutions that may not fit specific engineering problems or developing custom architectures from scratch – a time-consuming process requiring specialized expertise. Rescale significantly expanded AI Physics capabilities with comprehensive support for surrogate model development, providing engineers access to proven model architectures optimized for engineering applications alongside custom development frameworks. At SC25, the computing industry’s marquee event, NVIDIA announced Apollo, a family of open AI physics models that will be available on Rescale in 2026, further expanding the library of validated architectures. This flexibility allows teams to deploy validated models for common use cases or develop specialized architectures for unique requirements, dramatically accelerating project timelines while maintaining the accuracy traditional simulation demands.
Looking Ahead to 2026
As we get set for 2026, we’re committed to expanding the value we deliver to our customers and continuing to bring new capabilities to our users. We’ll be deploying production-ready autonomous agents at scale that enable true self-driving engineering workflows. AI Physics will broaden coverage across additional domains including predictive maintenance and digital twin applications, enabling organizations to leverage historical simulation data for operational intelligence. We’ll continue delivering novel compute offerings that make advanced capabilities more accessible, and we’ll ensure all platform capabilities deliver standalone value while working together seamlessly.
We’re grateful to our customers who partnered with us throughout 2025, providing feedback that shaped our roadmap and pushing the boundaries of what’s possible with AI-enhanced engineering. Together, we’re building the future of simulation-driven innovation.Explore all the exciting features we’ve introduced this year in our ongoing release notes. As we continue investing across all areas of the Rescale platform, be sure to check back regularly – there’s more to come as we enhance the customer experience and power industry innovation throughout 2026.