NovoLINC Targets AI Chip Cooling Bottlenecks with New Thermal Material

Built on a proprietary nanostructured architecture, the material is engineered to handle heat flux exceeding 100 W/cm². Beyond raw thermal performance, the product accommodates device warpage of up to 350 µm, allowing it to function across various interfaces, including direct chip-to-cold plate connections and heat spreaders. According to CEO Dr. Ning Li, reducing thermal resistance provides immediate headroom for higher chip performance and lower operating costs.

To meet projected demand from hyperscalers and server OEMs, NovoLINC is scaling its manufacturing footprint. The company is opening a new production and R&D facility in Sharpsburg, Pennsylvania, which will serve as the primary hub for commercializing the technology. Vice President of Operations Darren Goetz noted that the site provides the necessary infrastructure to support volume requirements as customers move from qualification toward full-scale integration. The company, a Carnegie Mellon University spin-out, currently maintains partnerships with the Open Compute Project and NVIDIA’s Inception Program.