Advancing Materials Innovation Velocity with AI
The rapid scaling of AI applications is placing unprecedented demands on semiconductor materials for advanced chips, compressing development timelines and restricting trial-and-error discovery. Conventional approaches to materials R&D — siloed data, manual knowledge retrieval, and low-throughput experimentation — are no longer sufficient to keep pace. Merck Electronics addresses this challenge through Materials Intelligence™ Solutions. By converging AI with materials science, we create a solution platform that integrates high-throughput process development, advanced metrology, and data-enabled AI to accelerate lab-to-fab learning.
In this talk, we will present an overview of our approach, including standardization of experimentation; an in-house large language model (LLM) for retrieval over proprietary reports and a materials knowledge graph to surface institutional know-how with provenance and access controls; Machine-learned interatomic potentials (MLIPs) enabled large-scale simulations; and Machine Learning (ML) guided experimentation that creates a Materials AI co-scientist to accelerate discovery toward sustainable, high-yield paths for next-generation AI hardware. Together, these capabilities close the lab-to-fab learning loop faster than any single tool can achieve alone — accelerating the delivery of next-generation materials for AI hardware at the speed the industry demands.
Key Technologies Covered
- Converging AI with materials science
- High-throughput process development
- Advanced metrology
- Data enabled AI
- Large language model (LLM)
- Machine-learned interatomic potentials (MLIPs)
- Machine Learning (ML) guided experimentation to accelerate materials discovery
