Some links on this page are affiliate links: if you buy through them we may earn a commission, at no extra cost to you.
Liquid cooling is becoming a mainstream design requirement for new, high-density AI and high-performance computing (HPC) data centers, but it has not replaced air cooling across the industry. For construction and facilities teams, the practical shift is that cooling infrastructure must be planned around the workload: high-density halls increasingly need liquid-ready layouts, while many conventional and existing data centers remain well served by air or hybrid systems.
That distinction matters. Liquid cooling is no longer an experimental add-on for the most demanding deployments, but a liquid-cooled server does not make a building liquid-ready. The decision reaches into plant capacity, piping routes, floor loading, redundancy, commissioning, maintenance access and heat rejection.
What “mainstream” means for data centers
Adoption is real, but the installed base remains mixed. In Uptime Institute’s 2025 cooling survey, 22% of respondents reported using direct liquid cooling, compared with 75% using perimeter air cooling. These are survey responses—not a census of every data center—and the categories are not necessarily exclusive. The figures show both sides of the transition: liquid cooling is in operation, yet air remains widespread. Uptime Institute’s 2025 survey
ASHRAE’s AI data-center framework describes direct-to-chip cooling as the dominant approach for AI and HPC environments, where rack densities are moving from roughly 120 kW toward several hundred kilowatts, with megawatt-class racks anticipated. That describes the direction of high-density design—not the average rack in every facility. ASHRAE’s integrated design principles
#1 Best Overall
- 【Speed Controllable】Easy Cloud axial fan 120v allows you to freely adjust the computer cooling fan speed according to your needs. This flexibility allows you to adjust fan operation to a level that best suits your environment, whether you require powerful cooling or a quiet work environment
- 【AC Plug】Dual-ball bearings have a lifespan of 50,000 hours. Easy Cloud small computer fan 120mm comes with 3V to 12V multi-speed controller, increases maximum axial fan speed and powers the muffin fan from an AC outlet. Just plug it into an outlet and start the 120mm pc fan
- 【Applicability】Designed to meet the cooling and ventilation needs of a variety of devices, including pcs, game consoles, appliances, entertainment equipment, solar equipment and more, this 120mm vent fan provides effective silent cooling and is also an ideal replacement for your existing 12v computer fan. No matter what type of equipment you have, this 120mm case fan ensures it stays at the right operating temperature, improving performance and extending life
- 【Parameter】120 x 120 x 25 mm ( 4.72 x 4.72 x 0.98 inches. ) | Rated Voltage: 12V | Airflow: 95.8 ±10M | Rated Current: 0.3A | Bearings: Dual Ball | Speed: 700RPM to 2800RPM | Power: 3.3W | Noise: ＜41dB
- 【Customer Support】We strive to offer the excellent services out of your expectations. If you have any problems with our product, please feel free to contact us at anytime
So the defensible conclusion is specific: liquid cooling is becoming a standard option, and often the preferred thermal architecture, for new high-density AI/HPC capacity. It is not yet the default for ordinary enterprise racks, every AI workload, or the existing data-center estate.
Why high-density computing is changing the building brief
The pressure comes from heat concentrated in a small footprint. High-powered GPUs and other accelerators can produce rack-level loads that are difficult or costly to manage with room air alone. Operators can extend air cooling with containment, higher airflow, colder supply air, in-row equipment or rear-door heat exchangers. Those measures remain useful, but they draw on finite room, power and mechanical capacity.
In Uptime’s survey, 68% cited higher rack density as an adoption driver. Sixty-three percent said direct liquid cooling becomes necessary above 20 kW per rack. Treat 20 kW as a survey-based view, not an engineering cutoff: a rack’s workable limit depends on server design, airflow, operating temperatures, heat capture, facility capacity and reliability targets. Some operators may support higher densities with air-side measures; others may choose liquid below that level to preserve headroom.
ASHRAE advises that future data centers account for the ability to add liquid cooling as IT power and cooling needs rise. For a construction project, that makes liquid readiness a planning issue even if the initial fit-out is mostly air-cooled: space, routes and plant decisions made early can be expensive to undo. ASHRAE’s overview of liquid cooling in mainstream data centers
Liquid cooling is a system, not a server accessory
“Liquid cooling” covers several architectures. The most important distinction is where the liquid collects heat and how the building disposes of it.
Rank #2
- Adjustable temperature control helps ensure optimal performance for rackmount such as network, server, music, and AV cabinets
- Noise controlled fans makes the cooling system useful for a quiet office or business space
- Compact design mounts to any 19" inch cabinet and takes up only 1 unit of space
- Simple and easy to use LCD display allows user to control temperature
- Air pumped through to the top exhaust system of the fan
Direct-to-chip cooling
In direct-to-chip, or direct liquid cooling (DLC), coolant flows through cold plates attached to heat-producing components such as processors and GPUs. A typical heat path is:
Chip → cold plate → server manifold → rack manifold → coolant distribution unit (CDU) → facility loop → heat-rejection equipment
The CDU is a key interface. It transfers heat between the IT-side technology-cooling system (TCS) loop and the facility loop, and commonly houses pumps, valves, temperature monitoring and controls. Details vary by system. Direct-to-chip is currently the most mature and broadly applicable liquid approach for AI/HPC, but it does not necessarily remove all air cooling: memory, storage, networking, power supplies and other components may still release heat into the room. ASHRAE’s data-center cooling guidance
Rear-door heat exchangers
A rear-door heat exchanger captures heat from server exhaust at the rack. The servers remain largely air-cooled internally, so this is a hybrid approach rather than a complete replacement for room cooling. It can suit a brownfield hall with elevated rack loads where a less invasive step is preferable to fitting cold plates throughout the IT fleet. It still needs liquid distribution and leaves some air-side cooling demand.
Immersion cooling
Immersion places servers or components in a dielectric fluid, in a single-phase or two-phase system. It can support high heat transfer and reduce server-fan demands, but changes hardware handling, service routines and technician requirements. Fluid compatibility, warranty support, fleet standardization and replacement workflows need to be settled before selecting it. It is a specialized option, not a universal next step.
Rank #3
- Thin Window Fan APPLICATION: Maximize Airflow with 120mm Fans, this mini window fan is very versatile and consume less energy, perfect for Cabinets, Server rack, Chassis, Plant, Mushroom Growing, Ice Fishing Shack, Chicken Coop, Generator Box and more
- Variable Speed Control: Small exhaust fan offers variable speed control for personalized cooling. It runs on AC power with versatile voltage options (110V-240V), fitting various regions. The cooling fan control governor is ideal for hard-to-reach spots, simplifying speed adjustments without unplugging. | Input: 100V-240V 50/60Hz Output: DC 3-12V 2A |
- Small Ventilation Fan: The fan features durable plastic and easy setup, reversible for DIY ventilation. It offers exhaust and intake for cooling stuffy spaces. This sturdy, adaptable fan is perfect for keeping your home cool and ventilated
- Dual-Ball Bearing: Brushless motors ensure a 50,000 hours lifespan for 24/7, allowing the fan to be positioned flat or upright with a wider heat dissipation area for maximum convenience
- PARAMETER of Computer Fan with AC Plug: 480 x 120 x 25mm ( 18.88 x 4.72 x 1in. ) | Rated Voltage/ Current: 12V 0.45A | Airflow: 108CFM | Speed: 3000RPM | Air Pressure (in H2O): 0.2 | Noise Level: 42 dBA ( All at full speed )
Liquid-to-air, liquid-to-liquid and heat rejection
Once heat enters a liquid loop, it still has to leave the building. A liquid-to-air arrangement rejects heat through air-cooled equipment such as dry coolers. A liquid-to-liquid arrangement transfers it to a facility-water loop, which may then use chillers, cooling towers, dry coolers or another system. Liquid cooling moves heat more directly from IT equipment; it does not make heat rejection disappear.
Do these 3 things before closing this tab:
1Repair Windows errors before they cause bigger problems2Scan for outdated or missing drivers - takes under a minute3Clear out junk files and repair common Windows errorsTrendForce has identified liquid-to-air designs as a transition path for sites without the water-distribution infrastructure required by some liquid-to-liquid systems. It estimated liquid-cooling penetration in AI data centers at 33% in 2025, up from 14% in 2024. That is an attributed market estimate for AI data centers, not a measure of all data centers or the global installed base. TrendForce’s AI data-center cooling estimate
Why adoption is accelerating—and what the evidence does not prove
Several developments point to a shift from pilots toward planned infrastructure:
- Operator experience: Uptime found meaningful direct-liquid use, while higher density was the most frequently cited adoption driver. At the same time, respondents identified lack of standardization (39%), cost (38%) and reliability concerns (35%) as barriers. Adoption is therefore not frictionless.
- AI platform design: Rack-scale AI systems are increasingly designed around liquid-cooled configurations. NVIDIA positions its GB200 NVL72 as liquid-cooled. The vendor also reports potential cooling-related savings in a particular comparison; such claims should be assessed against the stated baseline and system boundary, not treated as independent proof of savings at every site. NVIDIA’s explanation of its liquid-cooled platform
- Engineering guidance: ASHRAE has dedicated guidance on liquid cooling and an integrated AI data-center framework. These resources reflect the need to design cooling as part of the facility rather than treat it as a server option.
- A wider product ecosystem: Cold plates, CDUs, manifolds, TCS piping, controls and integration services are available from established suppliers. That gives buyers more components and design paths, but does not guarantee that products from different vendors will interoperate without qualification.
- Brownfield planning: ASHRAE guidance includes retrofit considerations, while Uptime respondents named retrofit ease as the leading viability factor (46%). The question is now not only whether new sites should use liquid, but whether and how existing buildings can support it.
Market growth, platform roadmaps and new product ranges are evidence of momentum. They do not mean the installed base has already transformed, nor that every workload labeled “AI” needs liquid cooling.
Where liquid cooling is a strong fit—and where air remains sensible
Liquid cooling is most compelling when high-density compute is central to the project: new hyperscale AI capacity, HPC installations, supercomputing, specialized cloud or colocation halls, and private AI clusters designed around powerful accelerator racks. In a greenfield project, the owner can coordinate the CDU location, piping, heat-rejection plant, service clearances and expansion plan from the outset.
The Tool Desk
Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Rank #4
- 【better after-use experience】 Temperature reduction provides an expected longevity extension and higher performance of a critical network component,These fans are overall very helpful for devices that get a bit hot and start to throttle down.
- 【choice of most users】It works great ,for DIY cooling fan or as an additional cooling ,fan for your gaming needs. like as router, cabinet, Modem, DVR, Receiver, Streaming ,boxes, x-box, SSD, Security Camera NVR, andriod box, stereo, T-Mobile gateway. Good balance of quiet and airflow. keeping electronics cool .Three specifications of fans, suitable for more usage scenarios .
- 【Custom shock absorbing feet】 four feet using environmentally friendly rubber, after testing, the softness of the feet that can smoothly grab the desktop, not too hard and desktop resonance .
- 【Fan parameters】Connecter: USB; Cable Length: 55cm Or 21 inches; Bearing type: Sleeve ; Life: 35000 hours / Dimension: 360mm(L) x 120mm(W) x 25mm(H) / 4.7x4.7x1 in. per fan; Rated Voltage:5V 0.2A; Speed: 1500RPM; Air flow: 56.7CFM; Noise:23dBA .
- 【Warranty & Packing List】Warranty: One-year quality assurance. Please contact us, If the product has any quality problems, it will be refunded within 90 days or replaced within one year | Packing list: A finished product .
Air cooling remains a practical choice for many conventional enterprise environments, lower-density general-purpose compute, mixed legacy estates and buildings without suitable liquid distribution. It can also be the rational choice where a retrofit’s construction scope and downtime outweigh the benefit of supporting a small number of denser racks.
The alternatives form a spectrum, not a binary switch. Containment, higher-capacity room or in-row cooling, rear-door heat exchangers, workload scheduling, equipment selection, a hybrid deployment, or relocating a workload to a liquid-ready facility may be more appropriate than a full direct-to-chip retrofit.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Why retrofits are harder than installing a CDU
A liquid-ready server can arrive at a building that is not liquid-ready. Before a retrofit, the team needs to establish whether the site can route and support the required loops, deliver suitable temperatures and flow, and reject the added heat. The design also has to work during maintenance and failure—not only at normal load.
Key checks include:
- Capacity and plant: Verify facility-water supply, flow, temperature, pumping and heat-rejection capacity. A CDU does not solve an undersized chiller, dry cooler or cooling tower.
- Routes and structure: Check risers, ceiling and underfloor space, penetrations, rack loading, floor loading and service access. Plan where CDUs and manifolds can be maintained.
- Water quality: Specify and monitor chemistry for the IT loop and any facility-water interface. “Water-cooled” does not mean ordinary tap water can be put through a cold plate. Poor chemistry or contamination can contribute to corrosion, biological growth, fouling and flow problems.
- Leak response: Design leak detection, isolation, drainage or containment, alarm routing and safe maintenance procedures. Confirm how a leak is isolated without creating an unacceptable thermal or availability event.
- Reliability and redundancy: Review pump and CDU redundancy, bypasses, failover behavior, power dependencies and response to loss of facility water. Define what happens to IT equipment if flow or temperature leaves its allowed range.
- Operations: Train facilities and server teams on connectors, coolant handling, maintenance, spare parts and escalation. Align commissioning, warranty terms and vendor service coverage with the operating model.
- Deployment constraints: Coordinate rack compatibility, server OEM qualification, shutdown windows, commissioning time and the treatment of adjacent air-cooled equipment.
A brownfield comparison should include more than one liquid design. Evaluate direct-to-chip conversion, rear-door heat exchangers, hybrid operation, a new liquid-ready hall, and workload relocation to a suitable colocation or hosted site. If piping, electrical, structural and controls work becomes extensive, a different location may offer better schedule or lifecycle economics.
Quick wins for a faster PC:
Scan for outdated or missing drivers - takes under a minuteDriver Scan →Clear out junk files and repair common Windows errorsFree Scan →Uptime’s finding that retrofit ease led viability considerations is a useful reminder: theoretical cooling performance is not the whole business case. The building, project schedule and operating team determine what can be delivered.
Best Value
- Pair of axial fans made to keep air flow and your equipment at low temperature
- Fits all standard 19” network cabinets; AC 110V Fan; 95/110CFM Airflow; 2600-2800rpm; 45dBA, Silent; AC cable 6.2ft and Ground wire 9" attached
- Network Cabinet Fan Applications - fan cooler panels, trays or server, media cabinets, computer case, DIY mount; overheat protection
- Steel Frame; Metal Finger Guard; Quick Mount Silicone Rubber Screws - Rivets; Self-tapping screws;
- Standard accessories exhaust replacement size: outer dimensions: 4.75”x4.75" - 4 inch between holes
Efficiency and water claims need a clear boundary
Liquid cooling can reduce fan energy and room-air-conditioning demand, and higher coolant temperatures can make chillers unnecessary in suitable designs. Dry coolers may reduce operational cooling-water use, and warmer loops can create opportunities for heat reuse. But none of these outcomes follows automatically from installing cold plates.
ASHRAE’s AI framework describes a warm-water, direct-to-chip case study that eliminated chillers, achieved a PUE near 1.10 and brought cooling-water use close to zero with dry coolers and limited adiabatic assistance. That is a particular case, not a guarantee or benchmark for every climate, load or design. ASHRAE’s AI data-center case studies and design principles
When comparing designs, ask what is included: CDU and pump power, chillers, dry coolers, cooling towers, room cooling, seasonal conditions, utilization and the baseline system. Water use should likewise be evaluated at the chosen boundary. A site may reduce direct cooling-water use while the electricity supply, coolant lifecycle or other parts of the system carry different resource impacts. PUE and WUE are useful measures, but neither alone describes total environmental performance.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
A practical architecture decision framework
| Project condition | Approach to evaluate first | Construction and operations question |
|---|---|---|
| New hall for rack-scale AI/HPC | Direct-to-chip, with the required facility loop and heat rejection designed in | Can the building support target rack loads, redundancy, service access and future expansion? |
| Existing hall with elevated density and mostly air-cooled servers | Rear-door heat exchangers or a hybrid air/liquid design | Can the project add liquid distribution without disproportionate structural, piping or downtime costs? |
| Mixed fleet at modest rack loads | Air cooling, containment or incremental upgrades; preserve liquid readiness if density may rise | Is liquid needed now, or is reserving routes and plant space the better investment? |
| Very high density with a standardized compatible fleet | Compare direct-to-chip and immersion against operational requirements | Can service, warranty, fluid handling and technician training support the chosen method? |
| Facility lacks a feasible retrofit path | New liquid-ready capacity or workload relocation | Is building work, schedule risk and ongoing service more costly than moving the workload? |
Before requesting proposals, assemble the facts that determine a design: current and target rack power; the server, accelerator, memory and networking mix; the share of heat expected to be captured by liquid; facility-water temperatures and available flow; heat-rejection capacity; required redundancy; project schedule; expansion plans; service and spare-parts needs; and OEM qualification and warranty requirements. Include whether the site must support air- and liquid-cooled racks side by side.
Then ask suppliers to state the interfaces, coolant and water-quality requirements, control and telemetry approach, redundancy assumptions, commissioning scope and failure response. Require savings claims to identify their baseline and whether pumps, CDUs and heat rejection are included. Industry guidance from ASHRAE and the Open Compute Project’s modular TCS guidance can help frame technical comparisons; site-specific engineering still governs the design.
What the transition is likely to look like
The near-term market is likely to remain mixed. Direct-to-chip is positioned to lead high-density AI/HPC deployments; rear-door and other hybrid systems offer paths for some existing halls; air cooling will continue to serve much of the conventional installed base; and immersion will remain an option for particular fleets and operating constraints. New facilities can make those choices deliberately, while retrofit projects must work around existing plant, routes and operations.
For construction teams, the most durable takeaway is not that every data center should become liquid-cooled. It is that high-density compute changes the facility brief. Owners should decide what density they are building for, what infrastructure can be added later, and how cooling, power, service and heat rejection will fit together before the design is locked.
Recommended Free Tools
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.
Leave a Reply