By Mark Toner
Australia’s data centre sector is experiencing unprecedented growth. Capacity is set to more than double from approximately 1350 megawatts (MW) in 2024 to 3100 MW by 2030.
Amazon, Microsoft and other hyperscalers are committing tens of billions of dollars to Australian facilities,
fundamentally transforming our digital infrastructure landscape.
Yet, amid the construction headlines, critical considerations remain around the post-commissioning operation and
maintenance of these facilities. The pace of development is outstripping the availability of skilled personnel, creating an operational capability gap that the facilities management sector must address.
A new era of complexity
Just 10 years ago, a one-megawatt facility represented significant capacity. Today, we are constructing 12+ MW data halls within 250+ MW campuses. Rack densities have progressed from five kilowatts (kW) to over 130 kW, while individual data halls have expanded from four MW to beyond 12MW.
Facility managers must now command specialist knowledge in high-voltage power distribution and high-density liquid cooling – capabilities that were peripheral to the role just a few years ago.
This densification also changes the scale of critical plants to maintain and operate, including chillers, generators, air handling units, heat exchangers, switchboards and power and cooling distribution systems. For example, a 100MW hyperscale facility could have 50 2MW chillers on its roof. A midsized commercial building would have two.
The liquid cooling imperative
The transition to liquid cooling represents one of the most significant operational shifts in the sector’s history, with propylene glycol solutions running directly to processor chips through precision engineered cold plates.
Immersion cooling systems employ specialised dielectric fluids that surround entire servers. Australia is now home to some of the largest liquid-cooled data centres globally, with international operators recognising these facilities as new benchmarks.
Facility managers are now expected to understand fluid chemistry – successfully maintaining concentrations and corrosion inhibitors within precise specifications to prevent pitting, scale formation and microbiological contamination.
The stakes are considerable: AI server racks can cost upwards of $3 million each, and the GPU chips they house are
both expensive and supply-constrained. Contaminated cooling systems risk clogging microfin heat exchangers, leading
to hardware damage and disruption to AI training workloads.
Energy and water efficiency
From 1 July 2025, data centres serving Federal Government workloads have required a minimum 5-star NABERS
energy rating. While technically limited to government contracts, experience suggests the broader industry will adopt
(or be mandated) with similar standards.
The business case supports this direction: efficient facilities attract quality tenants and improve operating margins. Power usage effectiveness (PUE) is the globally accepted energy usage metric for data centres. It is the ratio of total facility energy usage divided by IT equipment energy usage. A 5-star NABERS energy rating is equivalent to a PUE of 1.34.
Water consumption, meanwhile, requires greater attention. Many hyperscale facilities employ open-loop water systems
for enhanced free cooling, effectively trading energy consumption for water use.
Currently, this water is drawn from potable supplies. The cumulative demand from projected facilities would place significant pressure on municipal water resources. The industry has an opportunity – and arguably an obligation – to accelerate adoption of alternative sources such as recycled water, treated river water or desalinated supplies.
Bringing FM expertise upstream
One of the most effective interventions the sector can make is engaging facility management operators earlier in the
development cycle. Ideally, this should occur during detailed design, and certainly before commissioning.
Operational expertise adds substantial value to the design coordination process, ensuring facilities are maintainable and serviceable without compromising safety, uptime or operating cost. The facility manager’s perspective is likely to identify practical considerations that design teams can overlook.
The current pace of construction often limits facility manager representation in the design phase. Yet these are precisely the discussions where operational perspectives on water sourcing, maintenance access and system isolation
can shape improved long-term outcomes.
The data centre sector stands at an inflexion point. The construction pipeline is well-funded and accelerating, but
operational capability must keep pace.
Mark Toner is the Data Centre and Telecommunication Facilities Sector lead at A.G. Coombs
Image: Supplied
This article first appeared in the March 2026 edition of FM magazine. To download the entire edition for free, click here.
Read more: The HVAC&R skills gap won’t shift until we do more
