Predictive AI Reshapes Semiconductor Characterization
Semiconductor design teams are under pressure to keep up with rising process complexity, tighter margins and a growing number of process corners. Siemens’ Solido Characterizer responds with predictive AI that accelerates the creation of SPICE-based Liberty files, a foundational step in timing and power signoff. By combining an AI engine for multi-PVT Liberty generation with Solido LibSPICE, an AI‑accelerated characterization simulator, the software reduces Liberty file generation from weeks to days. This form of AI simulation software targets multi-corner, multi-format challenges, including LVF workflows, while preserving data quality and signoff‑grade accuracy. Integrated with Solido Analytics, Characterizer adds progress monitoring, automated reruns and quality assurance insights, helping teams keep characterization on predictable schedules. When paired with Solido Generator and Solido Fuse, it extends into generative and agentic AI workflows, enabling design cycle optimization that scales across libraries, IP blocks and design groups without overloading SPICE resources.
From Sequential Runs to Selective Simulation in Machining
In subtractive manufacturing, long and complex NC programs can stall verification and delay production. Hexagon’s latest NCSIMUL release introduces Selective Simulation to address that bottleneck, using GPU-accelerated engineering techniques to generate Rest Stock Previews during NC decoding. These intermediate stock models give programmers early visibility into part progression, allowing them to jump directly to critical operations instead of waiting for full sequential simulation. In a trial on a 47-hour mold machining cycle, a previously 48-minute simulation step dropped to less than two minutes before the target operation could be inspected. The result is faster iteration on toolpaths and fixtures, with verification keeping pace with modern programming workflows. Full NC code simulation with collision detection remains the final signoff step, but selective previews shift error detection earlier, reducing prove-out work on physical machines and tightening the overall design-to-production loop.
Compressing Design Cycles Across Digital Workflows
Both Solido Characterizer and NCSIMUL highlight a common trend: using AI and GPU acceleration to shorten feedback loops. In chip design, cutting Liberty file generation from weeks to days means library updates can keep up with new process nodes and format requirements, enabling more frequent design iterations without compromising signoff. In machining, GPU-powered Rest Stock Previews let programmers review problematic stages earlier, trimming hours from verification phases on complex parts. These advances in AI simulation software and GPU-accelerated engineering move verification from a blocking step to a near-real-time companion to design. The net effect is design cycle optimization: engineers can experiment more aggressively, validate more often and converge on manufacturable solutions faster. As digital twins and advanced simulators spread across disciplines, the distinction between design and verification increasingly blurs into a continuous, data-driven refinement process.
Integrating AI Simulators into End-to-End Digital Ecosystems
The latest tools are not standalone accelerators; they are designed to plug into broader digital product development ecosystems. Solido Characterizer feeds directly into the Solido Characterization Suite, where its baseline Liberty files train AI models in Solido Generator to create additional library views without rerunning SPICE. Coupled with Solido Fuse and Solido Analytics, it supports generative and agentic AI flows, automated quality checks and coordinated reruns, creating a tightly integrated characterization pipeline. NCSIMUL, meanwhile, operates as an NC program verification, simulation and optimization hub within a digital twin of the machining environment, linking G-code verification, stock evolution and process optimization under one roof. When these systems are connected to upstream CAD, CAE and PLM platforms, they enable seamless data exchange and traceability. This ecosystem approach ensures that gains in simulation speed translate directly into shorter time-to-market and more robust, repeatable engineering workflows.