Recommended Free Tools
Liquid cooling is often the better fit for racks with high-power processors or accelerators, but it is not automatically cheaper, more efficient, or more water-friendly than air cooling. The right choice depends on the workload, the whole cooling and heat-rejection system, the building and climate, water and energy constraints, and whether the site is new or being retrofitted.
How air and liquid cooling move heat
With conventional air cooling, server fans move heat from components into the room. Computer-room air handlers then transfer that heat to chilled water or another heat-rejection system. Air cooling can also use containment, careful airflow control, and ambient-air or water-side economizers to reduce mechanical cooling. Poor airflow management can waste energy through bypass air or overcooling.
Direct-to-chip liquid cooling moves heat from selected components into liquid circulating through cold plates. The liquid typically passes through a coolant distribution unit (CDU) and heat exchanger before the facility’s heat-rejection loop carries the heat away. The room usually still needs air cooling for components and equipment not connected to the liquid loop, such as memory, storage, power supplies, and network gear. ASHRAE describes hybrid air-and-liquid rooms as the norm outside full-immersion designs in its 2023 Handbook chapter on data centers and telecommunications facilities.
Immersion cooling puts compatible server equipment in dielectric fluid, capturing heat from more of the equipment than component-level cold plates do. It also changes how servers are configured and serviced. “Liquid cooling” is therefore not one interchangeable design: direct-to-chip cold plates, rear-door heat exchangers, and single- or two-phase immersion have different interfaces, service needs, and facility requirements.
#1 Best Overall
| Approach | Where heat is captured | What may still need air cooling | Key design consideration |
|---|---|---|---|
| Air cooling | Room air carries heat away from server equipment. | Air is the primary transport medium. | Airflow, containment, fan control, supply temperatures, and economization affect performance. |
| Direct-to-chip liquid | Cold plates capture heat from selected processors or accelerators. | Unconnected components and room loads often remain air-cooled. | CDUs, manifolds, hoses, facility piping, and compatible servers must work as a system. |
| Immersion liquid | Dielectric fluid surrounds compatible server equipment. | Room and facility loads may still require cooling. | Equipment compatibility and specialized service procedures are essential. |
When does a data center need liquid cooling?
There is no universal rack-density threshold established for when liquid cooling becomes necessary. High-density AI and high-performance computing (HPC) racks can increase heat flux until delivering enough air to the equipment becomes difficult, but the practical limit depends on the actual server configuration, rack thermal load, airflow design, and facility conditions. ASHRAE recommends designing around the rack’s real thermal load rather than relying on legacy assumptions about rack density.
Air cooling can remain appropriate for lower-density areas, while liquid cooling is assigned to high-density zones. ASHRAE’s AI data-center energy and thermal efficiency guidance recommends optimizing air management and economizers before adding advanced cooling, then matching liquid or liquid-assisted designs to the zones that need them. A hybrid design can be a transition during a server refresh or a durable solution for a site with mixed workloads.
Uptime Institute’s 2024 cooling survey asked operators when air cooling becomes too costly or inadequate and found differing respondent views, rather than an engineering cutoff. In that survey, 38% of respondents said they currently used direct liquid cooling and 49% said they did not use it but would consider it (Uptime Institute, 2024; n=453). Among direct-liquid-cooling users, 64% reported dielectric-cooled cold plates and 30% water-cooled cold plates (Uptime Institute, 2024; n=94; respondents could select multiple technologies). These are survey responses, not universal adoption rates or design limits. The survey identified increased cost, reliability concerns, maintenance, coolant leaks, supply-chain difficulties, and limited vendor choice as barriers. See the Uptime Institute 2024 Cooling Systems Survey.
Does liquid cooling use less energy?
It can, but the cooling method alone does not determine facility efficiency. Direct liquid systems add pumps, CDUs, piping, controls, and heat exchangers; the facility still has to reject the captured heat, and some residual loads may remain on air. Chillers, dry coolers, control sequences, server fans, and operating conditions all affect total energy use. Compare complete facility systems serving the same IT workload and reliability target, rather than comparing a liquid loop with an air handler in isolation.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Air cooling also has efficiency options. Containment, airflow management, fan control, supply temperatures kept within equipment limits, and economizers can reduce unnecessary cooling energy. ASHRAE’s efficiency guidance puts these fundamentals before choosing advanced cooling for high-density areas.
Power Usage Effectiveness (PUE) is total facility energy divided by IT equipment energy. A lower PUE indicates less facility energy overhead relative to IT energy; it does not by itself show water consumption, carbon intensity, server utilization, or useful heat recovery. Water Usage Effectiveness (WUE) measures site water use per unit of IT energy. DOE recommends considering cooling-tower evaporation and, where a fuller water-impact assessment is needed, water used indirectly to generate electricity. Definitions and federal data-center water-efficiency guidance are available from DOE FEMP.
Rank #3
DOE’s 2019 page reported PUE 1.06 and WUE 0.7 for the National Laboratory of the Rockies data center and its hybrid Thermosyphon Cooler Hybrid System. Those are results for that specific site and system, not performance guarantees for liquid cooling generally. DOE also notes that the design adds control loops that require an operations and maintenance plan. A lower PUE in one design should not be treated as proof of lower water use.
Can liquid cooling help recover useful heat?
Potentially. Direct-to-chip systems and warm-water loops can return heat at a higher temperature than low-grade exhaust air, which may make that heat more useful to another process. Recovery only has value when there is a nearby, steady heat customer and the supply temperature matches what that customer can use. Heat-reuse value should not be counted in a project comparison without a real sink and a workable connection.
ASHRAE’s integrated-design page gives a modeled example of over $4 million in annual savings for a 50 MW facility. That figure belongs to the page’s modeled scenario, not a general expected saving or a return forecast for another facility. The same page discusses integrated design principles at ASHRAE.
Rank #4
What does liquid cooling cost?
Capital costs can include cold plates or immersion tanks, CDUs, rack manifolds and hoses, facility piping, heat exchangers, leak detection, controls, commissioning, training, and server compatibility work. The scale and nature of the work differ substantially between a new facility designed around liquid cooling and a retrofit of existing racks.
A retrofit can require removing servers, changing chassis, discarding existing heat sinks, and reinstalling equipment. Labor, rework risk, and operational disruption can materially affect the economics. A California Energy Commission demonstration report described its supplier’s retrofit experience as economically unattractive, citing those kinds of installation and disruption costs.
For the Cab-cluster scenario in that 2024 report, estimated initial capital cost was $470,557.19, including $113,938 for facility modifications; modeled commercial-equipment pricing assigned $356,619.19 to the liquid cooling system before those modifications. The report estimated 348,663 kWh in annual energy savings and $39,154.85 in annual energy-cost savings for the scenario, using the report’s electricity-price assumption. These are project-specific demonstration and modeling figures, not current vendor quotes or market averages. The source is the California Energy Commission’s 2024 final project report.
Best Value
- Data Center Coolant
- 25% Inhibited Propylene Glycol
- JeffCool ISF 25
- High thermal conductivity
There is no broadly comparable market-average liquid-cooling premium or operating-cost reduction established here. A site-specific total-cost-of-ownership (TCO) comparison is more useful than applying a single premium or savings percentage. The Open Compute Project offers a TCO model for liquid-cooled data centers that compares power and cooling scenarios for new builds and retrofits.
Build a like-for-like TCO comparison
Use the same IT workload and reliability target in each scenario, and model local utility rates, climate, water tariffs, server refresh timing, redundancy, maintenance staffing, and expected utilization. Include:
- Server and cooling-system capital costs.
- Facility plant and electrical upgrades, installation, and commissioning.
- IT energy and cooling energy.
- Water and wastewater costs.
- Maintenance, spares, and staff training.
- Downtime and retrofit disruption.
- Any density or floor-space value the design actually enables.
- Heat-reuse value only when a real heat customer is available.
How to choose between air, liquid, and hybrid cooling
Evaluate the decision at the zone or workload level, not just as a facility-wide label. The factors below determine whether liquid cooling solves a real constraint or adds cost and complexity without enough benefit.
- Heat density and component power: Identify rack thermal loads and whether air can deliver cooling within the equipment’s limits.
- New build or retrofit: A new design can integrate servers, distribution, and heat rejection from the start; a retrofit must account for access, compatibility, labor, and disruption.
- Climate and water: Assess local ambient conditions, water availability, cooling-tower use, and the energy implications of the facility’s heat-rejection choices.
- Compatibility and vendor support: Confirm support for the server, cold plate or immersion configuration, coolant distribution, parts, and service procedures.
- Reliability and operations: Define leak detection, maintenance responsibilities, spares, control-loop ownership, and recovery procedures before deployment.
- Energy and water prices: Use local tariffs and the intended operating profile rather than generic savings claims.
- Heat reuse: Credit recovered heat only if a nearby user can take it consistently at a suitable temperature.
For many sites, the practical answer is not to replace every air-cooled rack. Optimized air cooling can serve lower-density areas, while direct liquid cooling or immersion can be evaluated for the zones whose thermal loads justify it. Compare each design using total facility energy, water, cost, reliability, and any credible heat-reuse benefit.
Quick wins for a faster PC:
Clear out junk files and repair common Windows errorsFree Scan →Scan for outdated or missing drivers - takes under a minuteDriver Scan →Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
