The rapid evolution of mainstream AI in the past 5 years has been astounding. Typically, AI experts align on a 5-level approach for GenAI, with Level 1 being Chatbots, Level 2 as "Reasoners," Level 3 as "Agents," and Levels 4 and 5 as "Innovators & Organizers." In 2024, Levels 1 and 2 were fairly achieved, Level 3 being the focus for 2025. While mainstream AI has achieved widespread adoption, applying it to complex and "high stakes" engineering and EDA workflows presents unique challenges.

This panel will discuss the key hurdles in adapting mainstream AI technologies to EDA, including tool integration, domain-specific fine-tuning, and ensuring "industrial-grade" reliability. We'll reflect on lessons learned from early AI (e.g., machine learning)—what has worked and what hasn't. Looking ahead, we will also discuss the current state of adoption of GenAI and Agentic AI for EDA, addressing whether it is more challenging and what capabilities and automation can power it.