Closed Cooling Systems

Closed cooling systems absorb and transfer heat from a facility.

Data centers contain servers, computers, and other electronic equipment to operate online services including cloud storage, artificial intelligence (AI) programs, and cryptocurrency mining (Leppert 2025, Offutt & Zhu 2025). Because this equipment generates heat, data centers require cooling systems to maintain a safe and consistent operating temperature (Singer et al. 2024, Kim & Lee 2024).

In a closed-loop cooling system, a liquid coolant runs through a system of pipes to take up heat from the server room (Veolia n.d.). This heated liquid then exits and passes through a heat exchanger, cooling it before it returns to the server in a continuous loop (Figure 1). This differs from an open loop system, in which the coolant leaves the system entirely after passing through the server room. Water is the most common liquid coolant (ClearWater 2025). Glycols are sometimes added to the coolant to lower its freezing point, though they can be toxic if the coolant leaks from the system.

Closed loop systems require less maintenance and water.

The components of the closed loop require less maintenance than an open loop system (Molander & Vedlugaite 2026, Evapco n.d.). The closed loop requires occasional replenishment to make up for gradual coolant loss. Water undergoes chemical pre-treatment to reduce buildup and corrosion in cooling pipes (Molander & Vedlugaite 2026, Cutler et al. 2018). Because the coolant is reused in a closed-loop system, it requires less pre-treatment than water for an open cooling system. Closed loop systems also have less buildup of contaminants that can lower heat transfer efficiency.

Some closed loop systems use direct liquid cooling, which applies coolant directly over the surface of Computer Processing Units (CPUs) and Graphics Processing Units (GPUs) (Torres Arango et al. 2025, Patterson et al. 2016). These systems dissipate heat faster than air cooling but are more expensive and require more equipment than a regular cooling system. The coolant is usually water or an aqueous polymer solution (Torres Arango et al. 2025).

Removing heat from the closed loop still consumes water or energy.

While the water used in closed loop cooling can dissipate heat more effectively than air cooling, the heat must be removed before the water returns to the servers (Ren personal communication 2026). The main methods for cooling the water loop are (Patel et al. 2026, Vertiv 2016):

• Water cooling, which consumes additional water through evaporative cooling.
• Air conditioner cooling, which consumes energy.
• Environmental air cooling, which uses outdoor air to remove heat more efficiently. The efficiency decreases as outdoor air temperature increases. Adiabatic-assisted cooling sprays water into the air to absorb more heat through water evaporation. This works best when the outdoor air is hot and dry.

As a result, data centers must choose one or a combination of these cooling systems based on the location’s climate and resource availability (Vertiv 2016). Data centers can switch between different cooling methods as weather conditions change.

For more information on data center resource demands, see: Data Center Water Use, Powering Data Centers, and Addressing AI Energy Demand.

Figure 1. Example schematic of a closed loop cooling system. A closed loop cycles cooling fluid through a facility like a data center to remove heat and into a heat exchanger. An external water supply takes the heat from the exchanger to cool the closed loop. Figure adapted from ChemStar Water n.d.

References

ChemStar Water (n.d.) Closed Cooling Water System: Guide to Efficiency, Protection, and Longevity. https://www.chemstarwater.com/closed-cooling-water-system-guide/

ClearWater Industries (2025) Should You Add Glycol to Your Closed Loop Glycol Cooling System? https://clearwatershelton.com/closed-loop-glycol-cooling-system/

Cutler D, Dean J, Daw J, Howett D (2018) Alternative Water Treatment Technologies for Cooling Tower Applications. National Lab of the Rockies. https://docs.nlr.gov/docs/fy19osti/71845.pdf

Evapco (n.d.) The Closed-Loop Advantage. https://www.evapco.com/closed-loop-advantage

Kim J, Lee H (2024) Computational Study of Thermal Performance Enhancement in High-energy Density Data Center Immersive Liquid Cooling. International Refrigeration and Air Conditioning Conference. Paper 2601. https://docs.lib.purdue.edu/iracc/2601/

Leppert R (2025) What we know about energy use at U.S. data centers amid the AI boom. Pew Research Institute. https://www.pewresearch.org/short-reads/2025/10/24/what-we-know-about-energy-use-at-us-data-centers-amid-the-ai-boom/

Molander I, Vedlugaite B (2026) System-Level Optimization of a Closed Loop Cooling System for Energy Efficiency: An Aspen Plus simulation of temperature programs, fouling effects and mitigation in plate heat exchangers and centrifugal pumps. Master’s Thesis. Lund University. 9239663. https://lup.lub.lu.se/student-papers/search/publication/9239663

Offutt MC, Zhu L (2025) Data centers and their energy consumption: frequently asked questions. Congressional Research Service. R48646. https://www.congress.gov/crs-product/R48646

Patel S, Kumar P, Liang Y, Kowalski T, Care N, et al. (2026) Prometheus: Toward Resilient Data Centers through Optimized Cooling Infrastructure. International Symposium on Computer Architecture. https://www.engineering.upenn.edu/~leebcc/documents/patel26-prometheus.pdf

Patterson MK, Krishnan S, Walters JM. (2016) On energy efficiency of liquid cooled HPC datacenters. Intersociety Conference on Thermal and Thermomechanical Phenomena in Electronic Systems. 2016: 685-693 https://doi.org/10.1109/ITHERM.2016.7517615

Ren S (2026) Associate Professor, Electrical & Computer Engineering Department, University of California, Riverside. Personal communication.

Singer B, Bingham DR, Corbett B, Davenport C, Gandolfi A, et al. (2024) GS SUSTAIN: Generational Growth — AI/data centers' global power surge and the sustainability impact. Goldman Sachs. https://www.goldmansachs.com/images/migrated/insights/pages/gs-research/gs-sustain-generational-growth-ai-data-centers-global-power-surge-and-the-sustainability-impact/sustain-data-center-redaction.pdf

Torres Arango MA, Torell W, Lena S, McGlocklin D (2025) 10 Ways to Harness the Energy and Water Efficiencies of Direct Liquid Cooling. Schneider Electric. White Paper 211. https://www.se.com/us/en/download/document/SPD_WP211_EN/

Veolia (n.d.) Closed Recirculating Cooling Systems. Handbook of Industrial Water Treatment. Chapter 32. https://www.watertechnologies.com/handbook/chapter-32-closed-recirculating-cooling-systems

Vertiv (2016) Freecooling, Evaporative and Adiabatic Cooling Technologies in Data Center: Applications in Diverse Climates within Europe, Middle East, and Africa. Vertiv White Paper. https://www.vertiv.com/495a36/globalassets/products/thermal-management/free-cooling-chillers/freecooling-evaporative-and-adiabatic-cooling-technologies-in-data-center-applications.pdf