Teratec / Nafems
▸ Simulation & Artificial Intelligence

AI in Simulation:
The Adoption Paradox

This report presents the findings of a consultation with leading experts, researchers and manufacturers in numerical simulation, run for the 2026 NAFEMS & Teratec event "AI & numerical simulation" on 10 June 2026. It looks at where AI deployment stands among the manufacturers who took part, the barriers they run into and the solutions now emerging.

Technical grid
01 — The Finding

From experiment to production: the missing link

85% of the organizations surveyed have begun testing or deploying AI in their simulation workflows. Of those, only 26% say they are satisfied operationally, a sign that scaling from pilot to production is still hard.

Adoption rate vs operational satisfaction
85%
Have started exploring AI
26.2%
Report operational satisfaction
Main barriers to adoption
Lack of usable data 51%
Moving from PoC to production 49%
Trust & validation of results 47%
Budget constraints 37%
Resources and skills 30%
Internal buy-in and prioritization 23%

Up to 3 answers per respondent

02 — The Obstacles

Training data is the biggest obstacle.

Almost half of respondents name the lack of usable data and the difficulty of reaching production as their main barriers to deployment.

Traditional architectures (GNN, PINN, ROM) struggle to scale to industrial meshes and start over on every project, with no way to reuse the data built up across earlier simulation campaigns.

03 — Emerging architectures

Transformer architectures are still underused.

Architectures built on Pre-trained Transformers cut the amount of data needed by relying on pre-training tailored to a specific area of physics. Even so, they are still very rarely used in industrial simulation workflows today.

5%
of the engineers surveyed have experimented with Transformer architectures in their simulation workflows.
04 — The application horizon

Toward hybrid models and production deployment.

Manufacturers are aiming beyond faster design tools. They want to move toward horizontal applications that can pool data from many different sources. The use case they rate highest is hybrid modeling, which combines simulation data with field measurements and feeds straight into production.

Applications most valued by industry
  • [01] Hybrid models (field + simulated data)
  • [02] Topology optimization
  • [03] Fast surrogate model
  • [04] Real-time AI simulation (CAD)

These applications are only feasible with AI-based simulators that can correlate heterogeneous data streams (3D displacement fields, IoT sensor time series, manufacturing metadata) which traditional tools cannot handle.