Hybrid Air and Liquid Cooling: A practical path forward for modern Data Centres

Data centres are changing faster than their cooling systems were ever designed to handle. High-density servers, GPUs, and AI workloads are pushing rack power well beyond traditional limits. For many facilities, air cooling alone is no longer enough. At the same time, moving completely to liquid cooling can feel risky, expensive, and disruptive. This is where hybrid air and liquid cooling comes in. It offers a realistic way to increase cooling capacity while continuing to use existing infrastructure.

What is hybrid air and liquid cooling?

Hybrid cooling combines conventional air-based cooling with liquid cooling technologies in the same facility. Instead of replacing air cooling entirely, liquid cooling is introduced only where it is needed most.

For example, a data hall may continue using CRAC or CRAH units for standard racks, while high-density racks are equipped with rear-door heat exchangers or direct-to-chip liquid cooling. The liquid removes heat at the source, and the remaining heat load is handled by the existing air system. In simple terms, air cooling does the broad work and liquid cooling handles the hotspots.

Server power density has increased steadily over the years. Traditional enterprise racks once operated at 3 to 5 kW. Today, it is common to see racks running at 15 kW or more, especially with GPU-based workloads. As rack density increases, airflow requirements rise sharply, leading to higher fan energy, noise, and cooling inefficiencies. Beyond a point, adding more air no longer solves the problem.

Liquid cooling is far more effective at removing heat, but a full conversion requires changes to piping, redundancy design, maintenance processes, and operations. Most organisations are not ready to take that step all at once.

Hybrid cooling has emerged because it allows operators to respond to higher densities without abandoning proven systems. It is a practical response to real-world constraints, not a technology trend.

While liquid cooling is highly efficient, an immediate full transition is not always the right choice. It involves significant upfront investment, changes to MEP infrastructure, staff training, and new operational risks. For existing facilities, this can also mean downtime and major retrofits. Hybrid cooling reduces this risk. It allows organisations to start small, validate performance, and build confidence over time. Instead of redesigning the entire data centre, liquid cooling is applied only where it delivers clear value.

How organisations can transition gradually using a hybrid approach

Start with assessment

The first step is understanding where the real problem lies. Not all racks need liquid cooling. Identify high-density workloads and racks that consistently run hot or limit future expansion. This assessment should include current rack power, airflow constraints, PUE, and growth projections.

Pilot liquid cooling at rack level

A controlled pilot is the safest way to begin. One or two high-density racks can be equipped with rear-door heat exchangers or direct-to-chip cooling. Rack-level solutions minimise plumbing changes and allow performance to be measured accurately. During this phase, teams can monitor temperatures, energy consumption, and operational impact without affecting the rest of the facility.

Expand in zones, not all at once

Once the pilot proves successful, liquid cooling can be extended to specific rows or pods that host high-density equipment. The rest of the data hall continues to operate on air cooling. This zoned approach keeps complexity manageable and avoids unnecessary upgrades to areas that do not need them.

Integrate operations and monitoring

As liquid cooling expands, monitoring becomes critical. Flow rates, temperatures, and pressure must be integrated into existing building management and DCIM systems. Operations teams should be trained gradually, starting with basic maintenance and emergency procedures. This ensures confidence and reliability as liquid systems scale.

Plan for the long term

Hybrid cooling does not need to be a temporary solution. Many data centres operate effectively with a mix of air- and liquid-cooled racks for years. For new builds or major expansions, lessons from hybrid deployments can inform whether a larger liquid infrastructure makes sense.

Key benefits of hybrid cooling

Hybrid cooling increases rack density without forcing a full infrastructure overhaul. It improves energy efficiency by reducing reliance on high airflow and excessive fan power. It also protects uptime by retaining familiar air-cooling systems as backup.

Most importantly, it aligns technical capability with business reality. Investment happens where it is justified, not everywhere at once.

Why system integration matters

Hybrid cooling is not just about adding liquid to racks. Poor hydraulic design, unclear responsibility between IT and MEP systems, and lack of monitoring can lead to performance issues. Successful hybrid deployments require careful coordination between cooling design, controls, and operations. This is where experienced system integrators play a critical role.

A practical path forward

For organisations facing rising densities and uncertain workloads, hybrid air and liquid cooling offers a balanced solution. It supports modern computing demands while preserving flexibility and control. Instead of asking whether to choose air or liquid, the better question is how to use both effectively. Hybrid cooling answers that question in a way that is technically sound, financially sensible, and operationally safe.

Not sure if liquid cooling is the right next step? Let Comfonomics evaluate your current setup and identify where hybrid cooling can add capacity without disruption.