AI-Accelerated Simulation Moves From Experiment to Baseline
Engineering teams are increasingly turning to AI-accelerated simulation to keep pace with rising design complexity and compressed delivery schedules. Instead of treating artificial intelligence as an add-on, software vendors are baking predictive AI and GPU-accelerated CAD tools directly into verification and characterization workflows. The goal is straightforward: engineering cycle time reduction without compromising accuracy or safety. In digital design, simulation workloads are exploding as more corner cases, formats and operating conditions must be validated. On the manufacturing side, long-cycle machining and multi-stage processes demand earlier insight into part state and process risk. Across domains, the emerging pattern is the same—use AI to prioritize what needs high-fidelity computation, and use accelerated hardware to render results fast enough for practical decision-making. That convergence is turning what used to be overnight or week-long runs into interactive, same-day iterations that better align with modern product development tempos.
Siemens Solido Characterizer Slashes Liberty File Generation Time
In semiconductor design, Siemens’ Solido Characterizer is a clear example of how predictive AI can compress critical workflows. The software targets SPICE-based Liberty file creation, a foundational step for timing and power analysis across process nodes. Traditionally, generating these libraries—especially with more corners, tighter margins and advanced formats like LVF—could stretch into weeks. Solido Characterizer uses a dual approach: a predictive AI engine to guide Liberty generation for multi-PVT and LVF workflows, and Solido LibSPICE, an AI-accelerated characterization simulator, to speed up SPICE-level analysis. Together, they reduce Liberty file generation from weeks to days while maintaining data quality. Integration with Solido Analytics adds quality assurance insights, progress monitoring and automated reruns, helping teams scale characterization across multiple IPs and design groups. The tool also feeds Solido Generator and Solido Fuse, enabling AI models to generate additional library views without full SPICE reruns, further shrinking turnaround within the Solido Characterization Suite.
Hexagon NCSIMUL Brings GPU-Accelerated Insight to Long Machining Cycles
On the shop floor, Hexagon’s latest NCSIMUL release shows how GPU-accelerated CAD tools can transform NC program verification. The new Selective Simulation feature uses GPU-accelerated Rest Stock Previews to build intermediate stock models during NC decoding. That gives programmers quick, visual snapshots of part progression in long and complex G-code programs. In a trial mold application with a 47-hour machine cycle, traditional sequential simulation demanded 48 minutes before the target operation could even be inspected. With Selective Simulation, Rest Stock Previews appeared in under two minutes, allowing earlier review of critical stages. Engineers can jump directly to operations that merit closer inspection, spot visible problems sooner and iterate toolpaths without waiting for full-sequence simulation. Final signoff still relies on complete NC code simulation with collision detection and material removal analysis, but the front-end review loop becomes significantly faster, reducing risk and improving throughput in high-value machining environments.
Toward Standardized AI and GPU Acceleration in Engineering Workflows
Taken together, tools like Solido Characterizer and NCSIMUL’s Selective Simulation highlight a broader transition: AI-accelerated simulation and GPU-driven previews are becoming baseline expectations, not niche extras. In digital design, predictive AI helps determine which SPICE runs, corners or library views truly require exhaustive analysis, while AI-accelerated simulators handle the rest with optimized performance. In manufacturing, GPU-accelerated previews turn long-cycle machining verification into a more interactive, staged process, aligning review checkpoints with real-world constraints. As more CAD and simulation platforms adopt similar patterns, engineering cycle time reduction becomes systemic, spanning from semiconductor characterization to NC program verification. The practical impact is fewer bottlenecks between design, verification and production, plus a greater ability to explore design alternatives without extending schedules. Over time, these capabilities are likely to integrate with broader generative and agentic AI workflows, further automating routine tasks while keeping engineers focused on high-value decisions.