AI and Digital Twins Transforming Power Device Innovation
Speaker: Steve Zamek, Director, Product Management, PDF Solutions
Abstract:
PDF Solutions draws on over a decade of experience working with leading power semiconductor manufacturers to examine how AI and advanced analytics can improve yield and reduce cost in SiC device manufacturing. Two production use cases are explored in depth: substrate-limited yield prediction, where machine learning models trained on substrate and epitaxial defect data are used to grade and disposition wafers before costly downstream processing; and predictive die binning for module assembly, where wafer sort data is used to forecast final test outcomes and optimize die disposition for maximum profitability. While both approaches have demonstrated meaningful results in controlled settings — including yield savings of several percent and dramatic cost reductions in assembly — the presentation is candid about the significant challenges encountered when moving from pilot to production, including inconsistent inspection data, poor traceability, missing data, and the gap between curated training datasets and real-world manufacturing data. The overarching lesson is that AI readiness begins with data readiness: manufacturers must consolidate data across silos, enforce traceability from raw materials to shipped modules, and standardize data quality before an AI strategy can deliver sustained value.
Keywords: Silicon Carbide, Yield Management, Predictive Binning, Data Traceability
Download the presentation today!
By downloading this information, I agree that PDF Solutions, Inc. may use my information provided herein to contact me about its products and services, even if I have previously unsubscribed from such communications. For information on our privacy practices and commitment to protecting your privacy, please review our Privacy Policy.