Materials become a bottleneck for AI infrastructure

AI’s rapid growth is turning advanced materials into a core constraint for semiconductors and data centers. Mike Finelli, chief technology and innovation officer and chief North America officer at Syensqo, says rising requirements for heat tolerance, purity, electrical behavior, chemical and plasma resistance, and long-term stability are pushing infrastructure beyond commodity materials. Syensqo is focusing on high-performance polymers, elastomers and specialty fluids that can support more powerful chips and denser data center architectures.

The company is developing materials for high-voltage data centers, sealing systems used in semiconductor fabrication, and cooling approaches such as dielectric fluids for direct immersion cooling. Finelli says work originally aimed at electric vehicles, including materials for high-voltage connections, thermal management and battery energy storage systems, can carry over to data centers as they contend with higher power density and reliability demands.

Sustainability is becoming part of performance rather than a separate goal. Syensqo assesses research projects through its Sustainable Portfolio Management tool and is working on heat transfer fluids intended to reduce environmental impact. The company is also using AI agents with Microsoft’s discovery tool to digitally synthesize molecular candidates, predict performance and sustainability traits, and select a smaller set for lab testing, creating a feedback loop in which AI helps build the materials needed to improve AI infrastructure.