AI-accelerated simulation tools like Ansys SimAI and digital-twin platforms built on PLM systems like PTC Windchill are compressing propulsion design-and-test cycles by predicting performance before physical parts are built, and by feeding as-built manufacturing data back into the next design iteration.
# Simulating Before You Build
Propulsion development for eVTOLs has historically followed a slow loop: design a part, build it, test it, measure the gap between predicted and actual performance, then redesign. Two categories of software are compressing that loop -- AI-accelerated simulation tools that predict physical performance without running a full physics-based solve, and digital-twin platforms built into product lifecycle management (PLM) systems that feed as-manufactured data back into the next design iteration.
## AI-Accelerated Simulation: From Physics Solvers to Learned Models
Ansys SimAI, part of the Ansys 2026 R1 release under Synopsys -- which completed its $35 billion acquisition of Ansys on July 17, 2025 -- is a metamodeling tool. It trains a predictive model on a set of prior full-fidelity simulation results, typically 30 to 100 runs according to Ansys's own product documentation, then estimates a new design's performance in seconds rather than the hours or days a conventional physics-based simulation takes. Ansys states the approach can cut overall design-cycle time by 10 to 100x, and each prediction carries a confidence score flagging designs that fall outside the range the model was trained on.
The 2026 R1 release ships SimAI in two forms: a cloud-based Premium SaaS product, and a desktop Pro version for teams that need to keep simulation data local. It arrives alongside generative tools like GeomAI for conceptual geometry and agentic add-ons that debug meshing and simulation-setup errors.
Two caveats matter for propulsion engineering specifically. Ansys's published SimAI use cases -- automotive crash testing, battery thermal management, power-inverter cooling, external aerodynamics, antenna design -- don't currently name eVTOL propulsion or aerospace motor testing. And a metamodel is only as good as the physics-based simulations it was trained on; it doesn't replace the underlying full-fidelity solve, it amortizes it across many design iterations.
## Digital Twins and the As-Built Feedback Loop
The second half of this trend runs through PLM software. PTC's Windchill platform, an established PLM system in aerospace and defense, has added what PTC calls an "agentic digital thread" -- AI-assisted workflows that connect data across design, manufacturing, and service so a change in one domain propagates to the others. PTC describes this as enabling closed-loop processes that connect design, manufacturing, and feedback, including automated generation of the manufacturing bill of materials from the engineering BOM, and process plans that bridge as-designed and as-built data.
For a propulsion program, that closed loop matters because certification-grade parts accumulate deviations between the CAD model and what actually gets built and measured on the shop floor: tolerances, material lot variation, assembly-specific quirks. A digital thread that keeps as-built data linked to the design record is what lets an engineering team feed real production data back into the next design pass, instead of treating manufacturing data as a one-way archive for quality audits.
## Where This Fits in Aerospace Propulsion Development
PTC names aerospace and defense as one of Windchill's core verticals, citing full lifecycle traceability, configuration control, and secure collaboration as the draw for regulated, mission-critical programs. That framing lines up with certification realities: FAA and EASA type-certification require traceable configuration control and documented substantiation for propulsion components -- exactly the kind of record-keeping a digital-thread PLM system is built to enforce.
Where AI-accelerated simulation and digital-twin PLM meet is earlier in the process: using a trained metamodel to explore motor, inverter, or battery-pack design variants quickly, then routing the design that gets built through a PLM system that tracks how the as-built part performed against prediction. That combination doesn't shorten certification testing itself, which stays governed by regulatory requirements for physical validation -- but it can reduce how many physical iterations a team needs before a design is mature enough to enter that process.
## The Net Effect
Neither Ansys nor PTC currently markets these tools with eVTOL-specific case studies, and it would be a mistake to read AI-accelerated simulation as a substitute for the physical, instrumented testing FAA and EASA certification requires. What the combination changes is the ratio of design iterations that happen digitally versus the ones that require cutting metal -- compressing the design-build-test-redesign loop earlier in the process, before a part reaches the certification test stand.
Synopsys (Nasdaq: SNPS) and PTC (Nasdaq: PTC) are publicly traded companies. Nothing in this article is investment advice -- it's a technical read on tools relevant to propulsion engineering workflows, not commentary on stock performance.
