CRAC Units Explained Precision Cooling Australia | Indigi Power & Cooling
CRAC Units Explained: The Complete Guide to Precision Cooling for Australian Data Centres
The data centre industry in Australia is undergoing rapid, sustained growth and the cooling systems that keep it running have never been more critical. As rack densities climb, AI workloads intensify, and the consequences of unplanned downtime grow more severe, precision cooling has moved from a back-of-house concern to a strategic infrastructure priority.
This guide covers everything facility managers, IT directors, and infrastructure engineers in Australia need to know about CRAC units: what they are, how they work, the differences between DX, CHW, and CDW systems, where they are deployed, how to maintain them, and how to select the right solution for your facility.
Why Precision Cooling Has Become a Business-Critical Priority
Australia's data centre market is one of the fastest growing in the Asia-Pacific region. According to IMARC Group, the Australian data centre cooling market was valued at USD 278.7 million in 2024 and is projected to reach USD 483 million by 2033, growing at a compound annual rate of 5.8%. The same analysis notes that Australian data centres already draw approximately 5% of the nation's total electricity around 1,050 MW a figure expected to rise to 8%, or approximately 2,500 MW, by 2030.
Cooling is the dominant reason why data centres consume so much energy. Research cited by IAEI Magazine finds that cooling systems account for 30–40% of total data centre power consumption, making them the second-largest electricity user in any facility after the IT equipment itself. Deloitte's technology predictions place the figure at 38–40% and project that global data centre electricity consumption could double to 1,065 TWh by 2030.
The thermal challenge is intensifying. Deloitte also reports that average rack density is anticipated to increase from 36 kW to 50 kW per rack by 2027 a 39% jump driven largely by AI and GPU infrastructure. Standard comfort cooling systems are not designed to handle this kind of concentrated heat load.
The consequences of inadequate cooling are severe. Equipment operating above its rated temperature fails prematurely, battery life in UPS systems degrades rapidly, and unplanned downtime can be catastrophic. Research from ABB reveals that unplanned downtime costs the typical Australian industrial business AU$349,000 per hour. For financial institutions, healthcare providers, and critical national infrastructure, that figure is even higher. Cooling is not a cost centre — it is a continuity mechanism.
What Is a CRAC Unit?
A CRAC unit Computer Room Air Conditioning is a precision-engineered cooling system designed specifically for the thermal management of data centres, server rooms, telecommunications infrastructure, and other environments where IT equipment operates continuously.
Understanding what a CRAC unit is, and why that definition matters, is the starting point for any informed purchasing or specification decision. For a detailed breakdown of the terminology and history of the term, see our dedicated article on CRAC unit meaning.
How CRAC Units Differ from Standard HVAC
The distinction between a CRAC unit and a standard split-system or commercial air conditioner is not simply a matter of brand or price it is a fundamental difference in engineering purpose.
According to TechTarget's data centre reference, CRAC units are generally more efficient than standard air conditioning systems and are specifically designed for medium to large data centres. The key differences include:
Temperature precision. CRAC units maintain room temperature to within ±1°C of setpoint. Standard comfort cooling systems cycle on and off and may vary by 3–5°C or more, which is inadequate for sensitive electronics.
Humidity control. Servers are damaged by both excessive humidity (condensation, corrosion) and insufficient humidity (electrostatic discharge). CRAC units actively control relative humidity typically to a range of 40–60% RH something standard HVAC units cannot do reliably.
Sensible vs. latent heat ratio. Standard air conditioners are designed to remove a significant proportion of latent heat (moisture) from the air, which is appropriate for human comfort. Computer rooms produce almost entirely sensible heat (dry heat from electronics). CRAC units are engineered with a high sensible heat ratio (SHR) of 0.9–1.0, meaning they focus on removing dry heat efficiently without unnecessarily drying the air.
Continuous operation. CRAC units are designed for 24/7/365 operation. Most commercial HVAC equipment is rated for intermittent use and will fail prematurely under the continuous loads of a production data centre.
Why comfort cooling fails in server rooms. A standard split system cooling a server room creates several risks: temperature stratification (hot spots near equipment racks), inadequate airflow volume, inability to maintain humidity setpoints, and mechanical failure from continuous operation. These failures do not happen gradually they cascade quickly, particularly in smaller rooms with high rack densities. For a comprehensive explanation of precision cooling technology, see our overview of CRAC cooling systems explained.
The Three Types of CRAC Systems
CRAC units are not a single technology. There are three primary refrigeration architectures in wide use across Australian data centres and server rooms, each suited to different facility sizes, heat loads, and infrastructure configurations.
DX CRAC Units (Direct Expansion)
DX CRAC systems use a refrigerant-based direct expansion cycle. A compressor pressurises refrigerant, which flows to an evaporator coil within the unit. Warm air from the room passes over the coil, heat is absorbed, and cooled air is returned to the space. Heat is rejected via a remote condenser (typically roof-mounted) or, in some configurations, a self-contained condenser.
DX systems are self-contained and relatively straightforward to deploy. They are well suited to small and medium-sized data centres, edge computing sites, telecommunications exchange rooms, and remote locations where installing a chilled water plant is impractical or uneconomical.
For a full technical explanation of how DX refrigeration works in precision cooling applications, see our article on DX CRAC units explained.
CHW CRAC Units (Chilled Water)
CHW CRAC units do not contain a compressor. Instead, they connect to a central chilled water plant that produces chilled water (typically 6–12°C) and circulates it through the building. The CRAC unit draws chilled water through a coil, passes warm room air over it, and returns the warmed water to the chiller for re-cooling.
CHW systems offer high cooling capacities, excellent energy efficiency at scale, and simpler in-room units with fewer mechanical components. They are the preferred technology for large data centres, co-location facilities, and enterprise environments where multiple CRAC units all draw from a central chilled water infrastructure.
The trade-off is infrastructure dependency CHW systems require a functioning chilled water plant, pipework, and pumps. Our detailed guide on CHW CRAC cooling systems covers CHW design, advantages, and specification considerations. Our CHW CRAC maintenance services page covers ongoing support for CHW-based facilities across Brisbane, Sydney, and Melbourne.
CDW CRAC Units (Condenser Water)
CDW CRAC systems sometimes referred to as water-cooled DX use a condenser water loop connected to a cooling tower or fluid cooler to reject heat. Unlike CHW systems where chilling is done centrally, CDW units contain their own refrigeration circuit and reject heat via the condenser water loop rather than to ambient outdoor air.
CDW systems are suited to very large data centres, hyperscale facilities, and environments where ambient outdoor temperatures make air-cooled rejection inefficient. They offer high efficiency and can achieve very low PUE in well-designed installations.
For a detailed comparison of condenser water cooling design and its applications, see our guide on CDW CRAC cooling systems.
DX vs CHW vs CDW: Comparison Table
This comparison is also covered in our broader overview of CRAC cooling systems explained, which provides additional context for technology selection.
CRAC vs CRAH: Understanding the Difference
CRAC and CRAH are frequently confused, and the distinction matters when specifying a system.
A CRAC unit (Computer Room Air Conditioning) contains its own refrigeration compressor and produces its own cooling through a direct expansion or DX refrigerant circuit. It is a standalone cooling device.
A CRAH unit (Computer Room Air Handler) does not contain a compressor. It is essentially a fan-coil unit that draws chilled water from an external plant, passes room air over the coil, and returns cooled air to the space. The cooling is generated by the chiller, not the air handler unit itself.
As a general guide, TechTarget notes that CRAC units are appropriate for loads below approximately 200 kW, while CRAH units connected to a centralised chiller plant are typically preferred above that threshold due to superior efficiency at scale.
In practice, many facilities described as "using CRAC units" are actually operating CHW CRAH units fed from a chiller. When specifying precision cooling, it is important to clarify which technology is actually being proposed.
Where CRAC Units Are Used in Australia
CRAC units are deployed across a wide range of sectors and facility types in Australia. Understanding the application context helps determine the right system type, capacity, and redundancy level.
Data centres and server rooms. The core application. From enterprise server rooms in a single building to multi-megawatt co-location facilities in Sydney, Melbourne, and Brisbane, CRAC or CRAH systems are the standard for precision thermal management.
Telecommunications infrastructure. Telco exchange rooms, radio base stations, and network operations centres require continuous precision cooling regardless of size. DX CRAC units are particularly common in distributed telco infrastructure due to their self-contained design.
Healthcare and pharmaceutical. Hospitals and pharmaceutical manufacturers operate critical IT systems and, in many cases, environmentally sensitive storage environments. Both precision cooling and humidity control are essential. Our pharmaceutical sector page details the precision infrastructure requirements in this space.
Banking and financial services. The uptime requirements in financial services are among the strictest of any sector. Tier III and Tier IV redundancy levels demand multiple CRAC units with N+1 or 2N configurations. Our work in the banking and financial services sector reflects this need for continuous, fault-tolerant operation.
Industrial and resources. Mining operations, process control environments, and industrial automation sites often house control room servers in challenging ambient conditions. Industrial sector cooling must account for dust, wide temperature swings, and remote site logistics.
AI and high-density computing. GPU clusters and AI training infrastructure generate heat loads that standard CRAC designs did not anticipate. As average rack densities rise toward 50 kW by 2027, the gap between conventional precision cooling and the demands of AI and high-density computing infrastructure is widening. Liquid cooling and in-row cooling are increasingly used alongside traditional CRAC for these environments.
CRAC Units and UPS Systems: Why Power and Cooling Must Be Engineered Together
One of the most common mistakes in data centre design is treating power protection and cooling as separate scopes of work. They are not they are deeply interdependent systems, and the failure of one can rapidly cause the failure of the other.
UPS systems, servers, and battery strings all generate heat as a by-product of their operation. Double-conversion UPS systems, which provide the highest level of power protection and are standard in mission-critical facilities, also generate the most waste heat due to their continuous AC-DC-AC conversion process. This heat must be accounted for in the room's total heat load.
Battery life is acutely sensitive to temperature. VRLA batteries the most common battery chemistry in UPS systems lose approximately half their service life for every 10°C above their rated operating temperature of 25°C. A room running at 35°C will have batteries failing at half their expected life cycle. Lithium-ion UPS systems offer improved temperature tolerance, but even lithium cells degrade faster in elevated ambient conditions. The importance of regular UPS maintenance is amplified when CRAC performance is unreliable, because temperature excursions may not be immediately visible.
CRAC failure in a room containing a UPS system creates a compounding crisis. The UPS continues to operate and generate heat; the batteries heat up; battery impedance rises and capacity falls; and if the situation is not caught quickly, the UPS may fail precisely when it is needed most. For high-temperature applications such as outdoor enclosures or industrial control rooms, high-temperature UPS systems are available but these are designed as a contingency, not a replacement for appropriate cooling.
The risks and costs of this kind of cascading failure are well established. ABB research places unplanned downtime costs for Australian industrial businesses at AU$349,000 per hour a figure that reflects both direct operational losses and downstream consequences.
The solution is an integrated design approach: treating UPS systems and CRAC systems as a single infrastructure scope. Our UPS & CRAC Design & Deployment service delivers exactly this thermal modelling, power load analysis, and coordinated commissioning of both systems to eliminate the gaps that lead to failures.
For those specifying UPS systems alongside cooling, our UPS buying guide and UPS power protection resources provide detailed guidance on technology selection, sizing, and redundancy design.
CRAC Installation in Australia
Installing a CRAC unit is not a standard air conditioning job. It requires specialist knowledge of data centre thermal dynamics, refrigerant handling, electrical integration, building management system (BMS) connectivity, and compliance with Australian standards.
Coverage and Compliance
Indigi Power & Cooling provides CRAC installation and commissioning services across Australia Brisbane, Sydney, Melbourne, and regional locations including remote mining and resources sites. We also support clients in Papua New Guinea.
All CRAC installations must comply with the relevant Australian standards and codes, including AS/NZS 3000 (Wiring Rules), refrigerant handling requirements under the Australian Refrigeration Council (ARC) licensing regime, and the facility's own operational and safety requirements. Equipment selection must also account for local ambient conditions an outdoor condenser or cooling tower in North Queensland faces very different design loads than the same equipment in Melbourne.
What a Professional Installation Involves
A thorough CRAC installation engagement begins well before any equipment is ordered. Key activities include:
Site assessment. Physical inspection of the room, including ceiling height, floor construction (raised floor or slab), existing electrical infrastructure, and available condenser locations.
Heat load analysis. Detailed calculation of IT equipment heat output, UPS losses, lighting, and personnel loads to determine total cooling capacity required and appropriate redundancy level.
Airflow design. Hot-aisle/cold-aisle containment, perforated tile placement (in raised floor installations), and airflow modelling to prevent hot spots and recirculation.
Equipment selection and procurement. Specifying the correct unit type (DX, CHW, or CDW), capacity, redundancy configuration, and BMS integration capabilities.
Installation and commissioning. Physical installation, refrigerant charging or water system connection, electrical connection, BMS integration, and functional testing under load.
Our installation and commissioning page provides an overview of our full project delivery capability across both UPS and CRAC systems.
CRAC Maintenance: Protecting Your Investment
A CRAC unit installed correctly will last 15–20 years under normal operating conditions but only if it is maintained properly. Deferred maintenance is the most common cause of premature CRAC failure, and the consequences extend far beyond the cost of the unit itself.
Why Preventative Maintenance Is Non-Negotiable
CRAC units operate continuously. Every component compressors, fans, filters, coils, humidifiers, expansion valves is under sustained stress. Unlike a comfort cooling system that runs intermittently, a data centre CRAC has no seasonal downtime during which problems become visible. Failures tend to emerge suddenly, often at the worst possible time.
Preventative maintenance identifies problems before they cause failures: a compressor running above rated current, a coil fouled with dust reducing heat transfer efficiency, a humidifier drain line partially blocked, or a fan bearing starting to wear. None of these issues is catastrophic on day one. All of them become catastrophic if left unaddressed.
Maintenance Schedule
Our CRAC unit maintenance services are structured around a tiered maintenance programme:
Monthly (site or remote monitoring): Filter condition check, temperature and humidity setpoint verification, alarm log review, refrigerant sight glass check.
Quarterly: Filter replacement, coil inspection and cleaning as required, drain pan and drain line inspection, electrical connection check, refrigerant pressure and superheat verification, fan belt inspection (belt-drive units).
Annual: Full mechanical inspection including compressor performance test, refrigerant recharge if required, humidifier element replacement, full electrical inspection, BMS setpoint calibration, and functional redundancy test.
For CHW-connected CRAC and CRAH units, system maintenance extends to the chilled water plant and pipework. Our CHW CRAC maintenance service covers the full system, including chiller plant servicing, across Brisbane, Sydney, and Melbourne.
Repair vs Replacement: The $5,000 Rule
At some point, every facility manager faces the question: is it worth repairing an ageing CRAC unit, or is it time to replace it? The costs of maintaining an old unit can accumulate quickly replacement parts, refrigerant charges, call-out fees while reliability continues to decline.
Our detailed guide on the $5,000 rule for CRAC repair vs replacement provides a practical framework for this decision, accounting for unit age, repair cost history, refrigerant type (legacy R22 units, for instance, carry significant ongoing compliance costs), and the risk premium of operating on ageing equipment.
As a general principle: if cumulative repairs over a 12-month period approach or exceed $5,000 on a unit more than 10 years old, replacement typically offers a better risk-adjusted return than continued maintenance. Modern units are substantially more efficient than equipment installed before 2015, meaning the energy savings from replacement often contribute meaningfully to the financial case.
Choosing the Right CRAC System for Your Facility
With three primary technology types and a wide range of manufacturers, capacities, and configurations available, selecting the right CRAC system requires a structured approach. The following factors drive the decision.
Room size and heat load. This is the foundational calculation. Total IT equipment power consumption (in kW or kVA) is the starting point, with additional loads for UPS losses, lighting, and personnel. The cooling system must be sized to match peak load, not average load, with adequate redundancy headroom.
Existing infrastructure. A facility with an established chilled water plant is a natural candidate for CHW CRAC or CRAH units. A single server room in a building with no existing cooling plant will typically be better served by DX. Infrastructure constraints are often the most significant driver of technology choice.
Redundancy requirements. The appropriate redundancy level (N, N+1, or 2N) depends on the criticality tier of the facility. Financial institutions and healthcare providers typically require N+1 or better. The cooling system must be designed as part of the overall redundancy architecture, not specified independently.
Efficiency and PUE. Power Usage Effectiveness (PUE) the ratio of total facility energy to IT equipment energy is the standard measure of data centre efficiency. According to the Uptime Institute's 2025 Global Data Centre Survey, the industry average PUE is 1.54, with the best-performing facilities approaching 1.1. Cooling system efficiency is the largest single lever available to improve PUE. Choosing a high-efficiency CRAC technology and maintaining it properly has a direct and quantifiable impact on energy costs and carbon footprint.
Airflow strategy. Hot-aisle/cold-aisle containment, raised floor plenum distribution, in-row cooling, and overhead supply are all valid airflow architectures depending on the room configuration. The cooling system and the airflow strategy must be specified together.
Emerging workloads. AI and GPU infrastructure is changing the density assumptions that underpin many existing CRAC installations. Perimeter air cooling the traditional CRAC placement model remains the most widely deployed approach: the Uptime Institute reports that 75% of facilities still use perimeter air cooling as their primary strategy. However, very high-density racks (above 20–30 kW) increasingly require supplementary in-row or direct liquid cooling. For facilities deploying AI and high-density computing infrastructure, a hybrid approach should be considered.
For guidance on the full range of technology types available, our CRAC cooling systems explained overview provides a useful reference. And because power protection and cooling are designed together in any mature infrastructure programme, our UPS buying guide is a natural complement to this resource.
Why Choose Indigi Power & Cooling
Indigi Power & Cooling (indigipc.com.au) is an Indigenous- and Veteran-owned Australian business specialising in integrated power protection and precision cooling infrastructure.
What distinguishes us from competitors in the market is not just what we supply it is the scope of what we do. The majority of companies operating in the CRAC and precision cooling space are either product-focused distributors or service-only contractors. Very few provide genuine end-to-end capability across both UPS and CRAC systems from a single, accountable team.
Our integrated approach means that when a facility is designed and deployed by Indigi Power & Cooling, the power and cooling systems are specified together, installed together, and maintained together. There is no gap between the UPS scope and the CRAC scope a gap that is precisely where most data centre failures originate.
Our service coverage spans Brisbane, Sydney, Melbourne, and regional Australia, including remote mining, resources, and industrial sites, as well as Papua New Guinea. We work across commercial, healthcare, pharmaceutical, financial services, industrial, and government sectors.
Our full lifecycle capability includes:
-
Design and specification heat load analysis, technology selection, redundancy architecture, and airflow modelling, delivered through our UPS & CRAC Design & Deployment service
-
Supply quality-assured equipment procurement across leading CRAC and UPS brands
-
Installation and commissioning professional installation to Australian standards; see our CRAC installation and commissioning and installation and commissioning pages
-
Preventative maintenance structured maintenance programmes via our CRAC unit maintenance services and UPS maintenance teams
-
Emergency response fast-response breakdown support when cooling or power events occur
To discuss your facility's CRAC requirements whether you are planning a new installation, reviewing an existing system, or managing ageing infrastructure contact our team or learn more about us.
Frequently Asked Questions
What does CRAC stand for?
CRAC stands for Computer Room Air Conditioning. It refers to precision cooling units specifically engineered for data centres, server rooms, and other environments containing continuously operating IT equipment. Unlike standard air conditioning, CRAC units maintain tight temperature and humidity tolerances 24 hours a day, seven days a week. For more on the origin and meaning of the term, see our article on CRAC unit meaning.
What is the difference between a CRAC unit and a standard air conditioner?
A standard air conditioner is designed for human comfort it cycles on and off, tolerates temperature variation of several degrees, has little to no humidity control capability, and is not designed for continuous operation. A CRAC unit is engineered for electronics: it maintains temperature to within ±1°C of setpoint, actively controls relative humidity, operates continuously without degradation, and is designed with a high sensible heat ratio to efficiently remove the dry heat produced by IT equipment. According to TechTarget, CRAC units are generally more efficient than standard air conditioning systems for data centre applications.
How long do CRAC units last?
A well-maintained CRAC unit typically has a service life of 15–20 years. Units that are poorly maintained, operated outside their rated conditions, or equipped with legacy refrigerants (such as R22) may require replacement sooner. Regular preventative maintenance is the single most important factor in maximising CRAC unit longevity.
How often should CRAC units be serviced?
CRAC units should receive at minimum a quarterly inspection and maintenance visit, with a comprehensive annual service. High-criticality facilities financial services, healthcare, co-location should also implement remote monitoring with regular alarm log review between physical visits. Our CRAC unit maintenance services page outlines our full maintenance programme structure.
What is the difference between DX, CHW, and CDW CRAC systems?
DX (Direct Expansion) systems are self-contained and use a refrigerant cycle to produce cooling they are suited to small and medium facilities. CHW (Chilled Water) systems connect to a central chilled water plant and are preferred for large data centres. CDW (Condenser Water) systems use a condenser water loop and cooling tower to reject heat, and are suited to very large or hyperscale facilities. See our full comparison in the DX CRAC units explained, CHW CRAC cooling systems, and CDW CRAC cooling systems articles.
Do CRAC units control humidity?
Yes. CRAC units include active humidification and dehumidification capability to maintain relative humidity within the range recommended for electronics (typically 40–60% RH). Standard air conditioners do not provide reliable humidity control and may cause either excess dryness (electrostatic discharge risk) or excess moisture (condensation and corrosion risk) in a server environment.
What temperature should a server room be maintained at?
ASHRAE (the American Society of Heating, Refrigerating and Air-Conditioning Engineers) recommends an intake temperature of 18–27°C for most IT equipment, with an expanded allowable range of 15–32°C for equipment rated to handle wider conditions. In practice, most Australian data centres target 20–24°C supply air temperature. CRAC units maintain these setpoints continuously standard air conditioning cannot do so reliably.
Can CRAC and UPS systems be installed together?
Yes and ideally, they should be. UPS systems and CRAC units share the same physical space and are thermally interdependent. UPS units, particularly double-conversion systems, generate significant heat that the CRAC must manage. Conversely, CRAC failure will degrade UPS battery performance and, eventually, cause equipment shutdown. Our UPS & CRAC Design & Deployment service treats both systems as a single integrated scope. See also our guide on what is a UPS for an overview of power protection fundamentals.
What is PUE and how does cooling affect it?
Power Usage Effectiveness (PUE) is the ratio of total facility energy consumption to the energy consumed by IT equipment alone. A PUE of 1.0 would be perfect efficiency (impossible in practice); a PUE of 2.0 means that for every watt consumed by IT equipment, another watt is used on facility infrastructure (cooling, power conversion, lighting). The Uptime Institute reports an industry average PUE of 1.54 in 2025. Because cooling accounts for 30–40% of total facility energy, it is the largest single contributor to PUE above 1.0. Selecting efficient CRAC technology, maintaining it properly, and designing effective airflow containment are the primary actions that improve PUE.
Should I repair or replace my CRAC unit?
The decision depends on unit age, cumulative repair costs, refrigerant type, and the efficiency of replacement options. Our detailed $5,000 rule guide for CRAC repair vs replacement provides a practical framework. As a general rule, if a unit is more than 10 years old and annual repair costs are approaching or exceeding $5,000, replacement typically offers a better financial and risk outcome particularly given the energy efficiency improvements in modern precision cooling equipment.
Summary
CRAC units are the thermal foundation of every reliable data centre, server room, and critical infrastructure environment. As Australia's data centre market continues to grow and as rack densities, AI workloads, and uptime expectations all intensify the quality of precision cooling has never mattered more.
For those managing existing CRAC infrastructure, the priorities are preventative maintenance, a clear repair-versus-replacement framework, and ensuring that UPS and cooling systems are managed as an integrated scope. For those specifying new installations, the technology selection DX, CHW, or CDW must be driven by site-specific heat load analysis, infrastructure constraints, and long-term efficiency goals.
Indigi Power & Cooling provides the full lifecycle of CRAC services across Australia: design and deployment, installation and commissioning, CRAC unit maintenance, and CHW CRAC maintenance integrated with our UPS maintenance and UPS brands Australia capability for a truly unified approach to critical infrastructure.
Contact us to discuss your precision cooling requirements.
Indigi Power & Cooling Indigenous & Veteran-Owned | Brisbane | Sydney | Melbourne | Regional Australia | PNG
