U.S. metal additive manufacturing company Fabric8Labs, in collaboration with the University of Illinois (UI), has utilized electrochemical additive manufacturing (ECAM) to produce pure copper cold plates for direct cooling of AI chips in data centers. The process employs liquid metal, enabling direct processing of pure copper rather than copper alloys for superior thermal conductivity. The research team used topology optimization to design pointed zigzag fin structures, improving cooling performance by 32% compared to traditional rectangular/cylindrical cold plates. If scaled, data center cooling power consumption could drop from 550 MW to 11 MW. The study is funded by the U.S. Department of Energy. Fabric8Labs secured $50 million in funding last November to expand domestic U.S. manufacturing capacity, transitioning from validation to mass production.