Smart Manufacturing CFA

Smart Manufacturing at SEMICON West 2027

AI in Semiconductor Manufacturing 
AI that has moved from pilot to production, with measured results.

• Agentic and generative AI for recipe management, recipe optimization and tool matching
• Virtual metrology, AI-enhanced FDC/APC and closed-loop, self-correcting process control
• Predictive maintenance, defect classification and faster yield learning
• Manufacturing-aware DTCO/STCO: feeding fab and yield data back into design choices
• Moving from pilot to scale, including retrofits of brownfield and 200mm fabs
• ROI scorecards covering downtime, scrap, cycle time, energy, and cost per wafer 

AI Orchestration, Digital Twins and Data Infrastructure  
The architecture and trust layer that connects AI, physical models and factory systems. 

• Semiconductor ontologies, knowledge graphs and multi-agent orchestration across MES and fab systems 
• Digital twins at tool, process, fab and supply-chain level, combining physics and ML models 
• Federated learning and privacy-preserving data sharing across fabs, OEMs and suppliers 
• Standard interfaces for AI and physical models and digital-twin vocabularies 
• Data governance, Human approval, explainability and decision authority for agentic AI 

Robotics, Physical AI and Human-Centered Autonomy  
Robots and people working together in cleanroom and fab environments. 

• Cobots, mobile manipulators and humanoid or dual-arm systems for fab operations 
• Custom end-effectors for wafers, dies, reticles and FOUPs 
• Edge AI, multi-function sensing and simulation-to-real training for robots 
• Robot–AMHS–tool interoperability and fleet orchestration  
• Robot safety, certification and governance for side-by-side human-robot work
• Autonomy levels set per task, AI copilots for technicians, and workforce reskilling 

Smart Manufacturing for Advanced Packaging and Heterogeneous Integration (AP/HI)  
Bringing frontend-level automation and data discipline to the backend.

• Tool automation and connectivity in backend and OSAT factories 
• Near-lights-out use models and factory designs for advanced packaging 
• Inline inspection, metrology and process control for hybrid bonding and die-to-wafer assembly 
• Unit-level traceability and chiplet genealogy from wafer to package to test 
• Test-data feed-forward and feedback, known-good-die strategies, and yield analytics for multi-die packages

IMPORTANT INFORMATION & 2027 DEADLINES
 

QUESTIONS?

Contact Anshu Bahadur
[email protected]

or 

Taylor Zhao
[email protected]

SPEAKER REQUIREMENTS

If your abstract is accepted and you are invited to present, please complete these requests:

• Speaker agreement
• Headshot and short bio (<500 words)
• PowerPoint slide presentation (16:9 format)
• PDF for distribution to attendees
• Identify a replacement speaker if unable to attend