AI in simulation: a new layer for engineering expertise
For decades, simulation engineers have been at the heart of industrial innovation. They translate physics into models, explore design options, and help companies make critical decisions long before anything is built.
Today, artificial intelligenceis entering this landscape. Not as a replacement for simulation, but as a new layer that can amplify engineering expertise. Yet adopting AI is not simply a matter of adding algorithms. It requires a shift in how engineers work with data.
Why AI should belong to engineers
In many organizations, models are built by specialists, trained on data that engineers may not fully control, and deployed as black boxes. This approach creates friction. Simulation engineers struggle to trust results they cannot explain, adapt or improve. Over time, AI becomes disconnected from real design constraints.
We believe the opposite approach is needed.
AI should be built by simulation engineers, using their own simulation data, their own assumptions, and their own understanding of the systems they design.
The emergence of the AI-builder
An AI-builder is not a data scientist stepping into an engineer’s role. It is an engineer who can design, train and refine AI models as a natural extension of their simulation work.
This evolution does not require engineers to become AI specialists. What it requires is the right foundation: tools and infrastructure that make AI accessible, reliable and consistent with engineering reality. When AI fits naturally into existing workflows, engineers can focus on what they do best: understanding systems, making trade-offs, and validating results.
For this to work, simulation data must be usable by design. It needs to be structured and qualified, reusable across projects, and accessible without long, manual preparation phases. Without these conditions, AI remains an attractive idea but a fragile reality. The gap between potential and practice simply becomes too wide.
Data as the missing link
Over time, simulation produces a wealth of knowledge. Each model, each result, each iteration captures decisions and assumptions that reflect real engineering expertise. Yet when this data is fragmented or locked inside tools, much of this knowledge becomes difficult to reuse.
When simulation data is organized and made ready for AI, its value changes. Engineers can build on past work instead of starting from scratch. They can compare designs more efficiently, identify patterns across projects, and train AI models that reflect their specific domain knowledge rather than generic assumptions.
In this context, AI is no longer an external system layered on top of engineering workflows. It becomes an extension of engineering judgment, shaped by experience, constraints and real-world understanding.
A change in responsibility and ownership
Turning simulation engineers into AI-builders also reshapes responsibility.
Engineers remain directly accountable for how models are built, which data is used, and how results are interpreted. This continuity is essential. It preserves trust in results, transparency in decision-making, and control over intellectual property.
Rather than shifting ownership away from engineering teams, AI stays in their hands. The teams who understand the systems best remain responsible for how AI is created and applied. This is what allows AI to scale in industry without becoming a black box—and what makes the role of the AI-builder both credible and sustainable.
Building the future of industrial simulation
As industrial systems continue to grow in complexity, the role of simulation engineers will keep evolving. AI will become a natural part of their toolkit, not a separate discipline.
At its core, this transition is about making better use of existing data and giving engineers the tools to build AI that reflects how they actually work.
This is where we see industrial simulation going. To support this direction, we recently completed a funding round, allowing us to move faster in building the foundations simulation engineers need to create, own and evolve their AI models.
Learn more about our funding round
Press release (english version)
🇬🇧 | Miura Simulation raises €2M to democratize the use of artificial intelligence in industrial simulation
Communiqué de presse (version française)
🇫🇷 | Miura Simulation lève 2 M€ pour démocratiser l’usage de l’intelligence artificielle dans la simulation industrielle (LinkedIn ↗).