What Is Indirect Liquid Cooling?

Indirect liquid cooling is a thermal management method in which a coolant fluid circulates through a system that is thermally coupled to heat-generating components but does not make direct contact with the electronics or other heat sources. Heat transfers from the component through a solid interface (typically a metal cold plate or a heat exchanger surface) into the coolant, which then carries that heat to a remote rejection point such as a chiller or cooling tower. The absence of direct fluid-to-component contact distinguishes indirect from direct immersion cooling approaches and makes indirect systems compatible with standard electronics packaging without requiring protective fluid coatings or specialized component enclosures.

Indirect liquid cooling draws from heat transfer engineering, fluid mechanics, and materials science. Liquid coolants carry heat far more efficiently than air: as reviewed by the Lawrence Berkeley National Laboratory's data center efficiency resources, the heat capacity of liquids is orders of magnitude larger than that of air, enabling the same volumetric flow rate to absorb significantly more thermal energy and thereby reduce the total fluid volume and fan power required for a given cooling load.

Cold Plate Systems

The cold plate is the primary interface component in most indirect liquid cooling deployments. A cold plate is a metallic block, typically copper or aluminum, containing internal channels through which coolant flows. It is mounted directly on the surface of a heat-generating component such as a CPU, GPU, or power module. Heat conducts through the metal base and into the coolant stream by convection. Channel geometry within the cold plate determines thermal resistance and pressure drop: microchannel cold plates with channel widths on the order of hundreds of micrometers offer high surface area for heat exchange and achieve low thermal resistance but require higher pumping pressure. As analyzed in a study of cold plate liquid cooling technology in Frontiers in Energy Research, single-phase cold plate systems using water or thermal oil can achieve cooling power of 12 to 35 kW per rack and can reduce data center energy consumption by approximately 35 percent compared to air-only cooling.

Cooling Loops and Heat Exchangers

Indirect liquid cooling systems use a closed-loop or open-loop circuit to transport heat from the cold plates to a rejection point. A cooling distribution unit (CDU) supplies conditioned coolant at a defined temperature and flow rate to a manifold serving multiple cold plates, collects the warmed return fluid, and routes it to a chiller or facility cooling water loop. Rear-door heat exchangers (RDHx) are an alternative arrangement, in which a rack-mounted heat exchanger positioned at the exhaust side of a server cabinet transfers heat from outgoing hot air to circulating cooling water, without requiring cold plates on individual components. These devices can remove 70 to 75 percent of the heat generated by a rack without modifying the servers themselves. For facilities seeking higher efficiency, waste heat recovered from the cooling loop can be redirected to district heating systems, a reuse pathway that the LBNL documentation identifies as a key efficiency benefit of liquid-cooled data centers operating with elevated coolant temperatures.

Performance Considerations and Thermal Design

Indirect liquid cooling system design involves balancing thermal resistance, coolant temperature, pump power, and leak risk. Coolant inlet temperature must remain low enough to maintain component junction temperatures within specifications, but higher coolant temperatures increase the potential for free-air or evaporative rejection without mechanical refrigeration, improving the facility's Power Usage Effectiveness (PUE). Corrosion inhibitors and material compatibility are engineering constraints: copper cold plates and aluminum manifolds require chemical management to prevent galvanic corrosion. The IEEE Spectrum analysis of liquid cooling in AI data centers identifies the rapid increase in processor thermal design power as the primary driver accelerating indirect liquid cooling adoption, with GPU-accelerated compute nodes regularly exceeding 700 watts per chip and exceeding air cooling capacity.

Applications

Indirect liquid cooling has applications across a range of high-heat-density computing and industrial contexts, including:

• High-performance computing clusters and AI training infrastructure
• Hyperscale and enterprise data centers managing dense GPU and CPU deployments
• Power electronics and inverters in electric vehicles and industrial drives
• Laser and radio frequency (RF) amplifier modules requiring precise thermal control
• Medical imaging equipment including MRI gradient amplifiers and CT scanner electronics