This Small Business Innovation Research Phase II project proposes to commercialize liquid submersion cooling for computer servers. Liquid submersion cooling involves submersion of heat-generating components in a non-electrically-conductive liquid, replacing air as the heat transfer medium. Liquid is significantly better than air to transfer heat, but historically has required cost-prohibitive capital expenditures due to the added complexity of previous liquid cooling architectures. The research objectives are to produce a system capable of being mass produced at low cost, and with compelling system features that drive customer demand.
The broader impact/commercial potential of this project includes lowering of one of the largest marginal contributors to US electricity use. The EPA estimates that data centers now use nearly 3% of US electricity, up from nearly 1% in 2000, with nearly half of power being driven by using air as the primary heat transfer medium. This high-efficiency system offers the potential to cut total energy use by nearly 50% by nearly eliminating energy for cooling and reducing server power through internal fan removal, while offering higher cooling performance and lower costs. Also, this new heat-recapture system offers the potential to eliminate nearly all server energy in many locales. Alternate cooling solutions that are cost effective only offer marginal improvements, and as computing becomes a larger part of the economy, the search for more energy and cost efficient technologies will become more critical.