OptiCool Technologies launched its new 120kW Rear Door Heat Exchanger (RDHx) – the highest-capacity rear door heat exchanger (RDHx) on the market. Capable of removing up to 120kW per rack, it’s purpose-built to meet the extreme thermal demands of next-generation AI and HPC workloads.
Engineered for seamless retrofit into existing environments, the new 120kW Rear Door Heat Exchanger – the highest-capacity unit of its kind – supports rack densities of up to 120kW and removes the need for separate zones or major infrastructure overhauls. Its proprietary two-phase refrigerant design, based on the thermodynamics of phase change, removes heat from the rack and returns it into the space at room neutral ambient temperature. This launch expands OptiCool’s existing RDHx lineup (30kW and 60kW), delivering a modular, high-capacity option for operators balancing density, uptime, and energy use.
“The 120kW Rear Door Heat Exchanger sets a new standard for cooling today’s high-density AI and HPC workloads,” said Matthew Roberts, VP of Sales at OptiCool Technologies. “No other rear door cooler comes close to this capacity. Our first-to-market, two-phase refrigerant design gives operators a direct path to cooling 80kW+ racks without the cost, risk, or vendor lock-in of other liquid cooling technologies.”
OptiCool’s Rear Door Heat Exchanger system requires significantly less energy than traditional air-based cooling approaches, helping operators reduce cooling costs and improve energy efficiency. The 120kW Rear Door Heat Exchanger leverages the same ultra-efficient design principles as OptiCool’s lower-capacity units, which have achieved power usage effectiveness (PUE) as low as 1.02 in tested deployments. This helps data centers advance ESG goals while preparing for next-generation compute demands.
Key Features:
- Market-leading capacity: Removes up to 120kW per rack, the highest of any RDHx available.
- AI/HPC ready: Supports the most demanding GPU clusters without specialized buildouts.
- Retrofit-friendly: Fits into existing environments with minimal disruption.
- Ultra-efficient: Cuts cooling energy use by up to 90%.
- Modular scalability: Grows with workloads from standard colo to extreme AI density.
