Data centres can be cooled using either air-cooled or waterless-cooled systems. Deciding which method is best suited for the future raises important questions. For instance, how much heat is produced by a data centre operating 24/7 to handle AI applications, cloud services, and real-time data streams? And how can this waste heat be efficiently removed without increasing overall energy consumption?
One thing is clear: stable IT operations are not possible without high-performance cooling. In an era of increasing computing power, higher server density, and growing sustainability demands, cooling is becoming a key strategic technology. It plays a crucial role in energy efficiency, system availability, and the future viability of data centres. At the same time, pressure is mounting on planners and operators to find solutions that can be integrated into both existing and new infrastructures in a cost-effective, scalable, and environmentally friendly way.
Waterless-cooled or air-cooled: in the end, which system can meet the growing demand? To answer this, the relationship between performance, efficiency, and operational reliability must be analysed and compared under real-world conditions. ESG criteria and total cost of ownership should also be included in the assessment.
In practice, air cooling remains the dominant technology in data centres, especially in existing buildings. Its popularity is no coincidence: it is considered mature, flexible to use, and comparatively cost-effective. However, with the increasing power density of modern IT infrastructures, air-cooled systems are increasingly reaching their physical and economic limits.
Classic air cooling is based on thermal convection: warm air rises and cool air is supplied. To effectively control this natural movement, operators often use so-called computer room air conditioning systems (CRAC) or computer room air handling units (CRAH). While CRAC units operate with an integrated refrigerant circuit, CRAH units use an external chilled water system to cool the air. In both cases, the hot air is drawn from the computer room, cooled, and recirculated.
These devices are usually installed along the sides or at the rear of the server room. A common practice is to combine them with a raised floor: cooled air flows upwards from the floor panels beneath the server racks and is drawn in by the airflow from the server fans. This principle is complemented by structural concepts such as hot and cold aisle containment. Server racks are placed in rows with the front (cold aisle) or rear (hot aisle) sides facing each other. This separation prevents air mixing, reduces energy losses, and improves cooling efficiency.
Free cooling systems (e.g. adiabatic or outside air utilisation), which do not require compressor operation at suitable outside temperatures, allow for further optimisation. However, air-based systems are still limited by their physical heat absorption capacity, especially at very high thermal loads.
Air-cooled systems offer economic advantages in terms of purchase and operation. They are often cheaper than water-based alternatives. The technical infrastructure, from CRAC units to air ducts, is widespread. As air cooling has been established for decades, there is wide availability with corresponding expertise. Spare parts, maintenance plans, and operational experience are available everywhere. Air-cooled systems are particularly suitable for retrofitting existing data centres. They can often be integrated or extended without major interventions. Additional air-conditioning units can be added, or the airflow routing can be optimised in a modular way.
The main weakness of air-cooled systems lies in their decreasing efficiency as power density increases. While conventional racks with 5 to 10 kilowatts of heat output per rack are easy to cool, these systems reach their thermal limits at 50 kW and above. In addition, the energy required for air movement increases considerably: fan energy use can account for up to 10% of total energy consumption.
Structural conditions must also be taken into account: Air cooling requires space for air ducts, corridors, and technical areas. The sound pressure level can also become a limiting factor with high air circulation, especially in peripheral environments or mixed-use buildings. Overall, it is clear that air cooling remains a proven solution for many classic applications. However, with increasing thermal loads, limited space, or ambitious efficiency targets, air-based systems are reaching their systemic limits and making room for alternative technologies such as water cooling.
With increasing power density and growing energy efficiency requirements, waterless cooling is becoming increasingly important in data centres. Unlike conventional air cooling, waterless cooling uses the higher thermal capacity of liquids to dissipate heat from IT components more efficiently. This allows for more precise temperature control and significantly reduces cooling energy consumption.
Waterless cooling is based on the principle of direct or indirect heat transfer through liquids. In direct liquid cooling (DLC), the cooling liquid is channelled directly to the heat-generating components, e.g., via so-called cold plates. These plates are heat conducting elements that are mounted on processors or other hot spots and transfer the heat directly to the liquid flowing through them.
Waterless cooling offers high cooling capacity, as liquids can transport heat much more efficiently than air. This enables the cooling of components with a high power density and reduces the risk of hot spots. It also offers good scalability, as waterless cooling can be adapted modularly to different requirements and is therefore suitable for different data centre sizes and configurations. Quiet operation is another advantage, with fewer or no fans required. This significantly reduces noise levels, which is especially advantageous in noise-sensitive environments.
In addition to the established methods of waterless cooling, immersion cooling also belongs to the category of liquid cooling technologies. In contrast to classic water cooling, the cooling medium does not circulate outside the IT components, but the servers are completely immersed in a special electrically non-conductive fluid. This method allows a particularly uniform heat dissipation and is especially suitable for high-performance applications with extreme thermal loads. Technologically, a distinction is made between single-circuit ("single-phase") and vaporisation ("two-phase") systems. Both variants have very low PUE values in practice and offer a high level of energy efficiency with low noise levels and minimal space requirements.
Despite the advantages mentioned above, the introduction of waterless cooling in data centres also brings challenges. The infrastructure must be adapted accordingly, which requires a higher initial investment. Safety aspects must also be taken into account, as the handling of liquids in the IT sector places special demands on impermeability and material compatibility. However, in many cases the long-term advantages outweigh the disadvantages, especially in terms of energy efficiency and operational reliability. Thanks to its high efficiency and scalability, water cooling represents an alternative to air cooling, especially in high-performance environments.
The table below shows a direct comparison of the strengths of air cooling and waterless cooling. Depending on the weighting of the criteria, the overview can provide an initial basis for decision-making.
.png?width=3433&height=1689&name=Blogbeitrag_Tabelle_Kuehlungssysteme_EN%20(1).png)
The increasing energy density of modern data centres, mainly due to high-performance computing (HPC) and AI applications, places new demands on cooling technologies. At the same time, environmental, social, and governance (ESG) criteria are becoming the focus of investors and regulators. This makes water-cooled cooling increasingly the preferred solution.
Hyperscalers such as Amazon, Google, and Microsoft are increasingly investing in waterless cooling technologies to make their data centres more efficient and sustainable. Microsoft, for example, has installed the aquatherm blue plastic piping system in several data centres, which stands out for its durability, corrosion resistance, and high energy efficiency. Novva's hyperscale data centre, Credit Suisse's data centre, and DuPont Fabros Technology's data centres also use aquatherm piping systems.
Therefore, waterless cooling is not only a solution to the technical challenges of modern data centres but also plays a significant role in meeting ESG criteria and implementing green IT initiatives.
The switch from air cooling to waterless cooling in data centres involves more than just a change in cooling technology; it also requires an adapted infrastructure. Piping systems play a crucial role, forming the physical connection between the central cooling components to create an integrated system. This system spans from the heat exchanger, which absorbs waste heat, to the pump lines and the cold plates or chassis-level coolers in the servers.
In practice, the quality of the piping system is crucial for efficiency, operational reliability, and maintenance costs.
The cooling concept can only achieve its full potential if the piping system transports the fluid reliably and with minimal losses. Leaks, pressure losses, or heat buildup in pipelines directly affect operational safety and, in the case of critical infrastructures, the availability of entire applications.
With the aquatherm blue pipe system made of PP-RCT (polypropylene random copolymer), aquatherm offers a solution tailored to the specific requirements of data centres. Compared to classic steel or copper pipes, PP-RCT has impressive material properties that optimise the entire life cycle of a cooling system. The advantages of aquatherm pipe systems made of PP-RCT compared to metal pipes are manifold.
Corrosion resistance
The plastic is resistant to oxidation, deposits, and electrochemical corrosion, without inhibitors.
Reduced weight
This facilitates transport, handling, and installation, especially in large-scale projects or complex installations on raised floors and roof areas.
Fusion connection
Pipes and fittings are thermally fused, without adhesives, joints, or flames. The result: a homogeneous, leak-free connection with maximum operational reliability.
Hydraulic flow efficiency
The smooth inner surfaces of the PP-RCT pipes minimise friction losses in the cooling circuit. This reduces pressure loss, allowing efficient and even distribution of the coolant throughout the system.
Minimised heat loss
The PP-RCT material structure has a significantly lower thermal conductivity than metallic materials. This means that temperature losses along the entire length of the pipe are kept low and the overall efficiency of the cooling system is optimised.
Modularity and scalability
Thanks to aquatherm's precise factory prefabrication, aquatherm systems can be tailored to project requirements, with shorter installation times and fewer sources of error on site.
Certifications and sustainability
aquatherm blue complies with international standards (ISO 9001, ISO 14001, ISO 50001) and is documented in an EPD. Therefore, the choice of pipe material is not just a technical detail but a strategic decision that impacts efficiency, safety, and sustainability in operation, particularly in terms of ESG reporting and green technologies.
The choice between air and waterless cooling is not simply a decision between old and new technologies, but rather one based on differing requirements, conditions, and objectives.
Air cooling continues to provide a reliable solution for applications with moderate thermal loads. It is cost-effective, easy to integrate into existing buildings, and remains a viable option for data centres with low to medium power density.
Waterless cooling, on the other hand, specifically addresses the challenges of modern IT infrastructures. It is ideal for environments with high rack density, rapid growth, or ESG-driven sustainability goals. With superior energy efficiency, low noise levels, and enhanced scalability, it is the most cost-effective long-term solution, particularly for new builds or expansions of hyperscale environments.
Hybrid cooling systems offer a middle ground; however, due to the combination of technologies, they present complex requirements in terms of structural equipment, control, monitoring, and maintenance.
It is crucial to adapt the cooling strategy to the specific context, considering technical, economic, and regulatory factors.