AI Meets Formal: From Trust Anchors to Productivity Gains

AI promises significant productivity gains in hardware design and verification, but hardware development demands strong guarantees. "Untamed" AI agents cannot provide the level of trust needed for safety-, security-, and cost-critical silicon. In this talk, we argue for a pragmatic middle ground, harnessing AI for speed and scale while keeping correctness anchored in deterministic, well-defined processes. We outline a spectrum of approaches, from formally anchored verified-by-construction AI-assisted design flows to AI-enhanced verification that operates within well-defined guardrails. The goal is not to replace rigor with probability, but to use AI to automate tasks without sacrificing determinism, traceability, or trust.

Dr. Tobias Ludwig is the Founder and CEO of LUBIS EDA, a company dedicated exclusively to formal verification for digital IP and SoC teams. With more than a decade of experience in formal verification, Tobias has scaled LUBIS EDA from an early-stage startup into a focused formal verification company serving complex silicon teams worldwide, bringing formal methods into practical, day-to-day sign-off for high-risk designs. Tobias holds a PhD in Computer Engineering with a focus on formal methods and hardware verification.