JEEVAN SHANBHAG

Title of the Talk :
Agentic AI in the Fab: Coordinating Recipes, Equipment, and Control Systems at Scale

Abstract of Talk:
Semiconductor manufacturing is among the most complex cyber-physical systems in operation, requiring precise coordination across heterogeneous equipment, control software, and fab-wide networks under strict real-time, safety, and reliability constraints. As the industry evolves from predictive analytics toward autonomous operations, agentic AI—systems capable of observation, reasoning, communication, and action—emerges as a critical architectural shift.
This keynote examines the role of agentic AI in coordinating semiconductor recipes, equipment behavior, and control systems at scale. Unlike conventional AI deployments, manufacturing autonomy demands long-horizon decision making, deterministic execution, low-latency edge computation, and robust communication across distributed control layers. Drawing on industrial experience, the talk highlights key challenges in multi-agent coordination, authority hierarchies, semantic consistency, and the secure execution of AI-driven actions.
The keynote argues that communication reliability and Zero-Trust execution are foundational to trustworthy autonomy and positions semiconductor fabs as a reference architecture for designing scalable, secure, agentic AI systems applicable to broader critical infrastructure domains.