Syensqo said it is using AI agents to digitally synthesize and rank millions of molecular candidates, narrowing them to about 100 for lab testing to speed materials development for AI infrastructure.
The push comes as AI workloads drive semiconductors and data centers toward physical limits in heat, voltage, purity, chemical resistance and long-term reliability, making advanced materials a tighter bottleneck.
Syensqo is targeting high-voltage data center architectures, semiconductor-fab sealing materials and direct-immersion cooling fluids, while adapting technologies first developed for electric vehicles and battery storage.
Sustainability is being built into research from the start, with the company saying 88% of its portfolio qualifies as sustainable and aiming to cut the environmental impact of next-generation heat-transfer fluids.
The company sees a feedback loop in which AI helps create better materials for AI hardware, which then improves the computing systems used to accelerate further materials discovery.