"With MODSIM, we can cut a design that takes 40 hours to 4 hours, and if we add AI, we can do it in 4 seconds."

Michelle Ash (미쉘 애쉬), CEO of Dassault Systemes' simulation solution SIMULIA, said MODSIM and AI are redefining engineering by running modelling and simulation on a single platform. She stressed it can deliver faster, cheaper products to more people while reducing environmental impact.

On June 11, she held a media briefing at the SIMULIA User Data 2026 conference hosted by Dassault Systemes Korea. She highlighted changes MODSIM and AI integration will bring to engineering and presented 3 keywords.

First is Physics Foundation, which is focused on improving speed.

Ash said it is boosting speed by redesigning the solver engine, which is core to simulation, on a GPU basis in cooperation with Nvidia. She said they have confirmed computing speed is clearly getting faster.

Next is MODSIM. Ash said CAD design, mesh setup and simulation analysis were previously carried out sequentially, but MODSIM allows users to check simulation results immediately at the design stage. She said Ford adopted MODSIM and cut design time from 40 hours to 4 hours. The last is AI. She said AI cuts the shortened 4 hours down to 4 minutes, adding that MODSIM is the first step for using AI.

Dassault Systemes is emphasising '3D Universe' as a long-term vision based on virtual twins.

She said 3D Universe supports simulation of the full life cycle in a virtual environment, from product design to manufacturing, use, maintenance and recycling. Ash said understanding the entire product life cycle before making a prototype can improve time to market, customer satisfaction and environmental impact.

The technology base supporting 3D Universe is the Virtual Twin, the virtual companion and Generative Experience, an integrated environment that combines the Virtual Twin and virtual companion.

The Virtual Twin focuses on simulating how a product will operate in the real environment based on the laws of physics. AI also plays a core role there. Ash said it plans to launch a 'Virtual Twin of Physics Behavior' combining AI and machine learning in July this year.

Virtual companions are agentic AI-based virtual assistants offered in 3 versions: Aura, Leo and Marie.

Aura handles business tasks such as project planning and supplier analysis, while Leo supports mechanical, simulation and fluid engineering. Marie plays the role of a scientist helping develop new materials and products.

Ash demonstrated a process using Leo to generate a 3D CAD model from a single photo, conduct a torsion test and then review part fit. She said some virtual companion functions will be available from July and the rest will be rolled out sequentially by the end of this year.

She also added that SIMULIA AI uses large language models from various companies such as OpenAI, Mistral and Anthropic in ways suited to their purposes. She said it adds simulation domain knowledge accumulated over the past 40 years through collaboration with customers.

Ash said virtual companions will lower the barrier to entry for simulation and bring major changes to expert roles. She said experts will use virtual companions to validate tools so non-experts can use them well, set up workflows or create guardrails. She said non-experts will handle low-risk activities, while experts will handle tasks such as verification and certification.

Dassault Systemes announced cooperation early this year focused on building an industrial AI platform by integrating the Dassault Systemes virtual twin platform with Nvidia AI computing infrastructure. Ash said the SIMULIA unit's cooperation with Nvidia is proceeding on 2 fronts: speeding up the solver and AI.

She said it is rewriting the solver for GPUs using Nvidia's CUDA technology and is also pursuing integration between Nvidia's digital twin solution Omniverse and the Dassault Systemes platform.