Why AI data centres are becoming a water planning test

Australia's AI boom is often discussed in terms of computing power, productivity and economic growth. However, behind every new data centre is a more practical question for the water sector: can local systems see, understand and plan for the demand before it arrives?

Australia is currently home to around 250 to 300 data centres, with dozens more in planning or development. The sector is growing rapidly and so is its resource footprint.

Data centres now underpin cloud computing, banking, cybersecurity, health systems, research, artificial intelligence and the digital services we use every day.

These facilities can create large, concentrated and continuous demand on local water systems, particularly where water is used for cooling. That demand can emerge quickly, often in locations where utilities are already contending with climate variability, leakages, drought planning and long-term supply security.

Cooling choices can significantly change water demand profiles. Some systems minimise water use but require more electricity, while others reduce energy intensity but consume more water. For utilities, these design decisions can materially influence local servicing needs.

Data centres are becoming a new test of water planning readiness, requiring earlier engagement, better visibility and coordinated planning to support sustainable growth.

A different kind of water demand

Data centres behave differently to many traditional water customers with established planning pathways and known operating patterns.

A typical small one-megawatt data centre using traditional cooling methods could consume approximately 25–26 million litres of water each year. However, their water demand is heavily shaped by design choices including cooling design, compute intensity, local climate, operating model, water source and efficiency measures.

It means two data centres of similar size may create very different impacts for local networks. For instance, whether they use recycled water and efficient cooling technology or rely primarily on potable water and evaporative cooling.

Cooling technology also influences when water demand occurs and how seasonal peaks intensify, with implications for both water system planning and energy use during summer periods.

This variation matters because utilities need early visibility of likely demand profiles to make informed, confident decisions.

The visibility gap

Water utilities are experienced infrastructure planners; however, the challenge in data centre growth is the combination of speed, scale and uncertainty.

A utility may be asked to assess significant new demand before it has a full picture of a facility's cooling design, water source, ramp-up timing, peak demand scenarios or drought contingency measures. Without early clarity on cooling pathways, utilities may be planning blind to a key driver of future water demand. If demand is underestimated, systems can come under pressure.

If demand is overestimated, capital can be allocated too early or in the wrong place. If demand arrives faster than expected, reactive investment decisions are more likely.

This is where better data becomes essential.

Water utilities need earlier visibility of proposed demand, likely operating profiles and servicing options to understand how the scaling of data centres will affect local supply, storage, treatment and distribution infrastructure.

Policy signals and water accountability

The national policy direction is becoming clearer.

The Australian Government's Expectations of data centres and AI infrastructure developers (March 2026) signals a clearer direction for the sector, calling for innovative and efficient water use, early engagement with utilities and communities, and consideration of local conditions in cooling design and water sourcing.

Notably, Expectation 3 – 'Use water sustainably and responsibly' – calls on developers to avoid placing upward pressure on local water resources, a principle that speaks directly to the concerns of water utilities planning for long-term supply security.

This reflects a broader shift in infrastructure expectations. Access to water alone is no longer enough. Efficiency, resilience, transparency and community confidence are increasingly central to project viability.

For water utilities, this is an opportunity to engage early. Water efficiency should be assessed alongside broader infrastructure impacts, particularly where lower water use may increase electricity demand during peak periods.

Australians understand the value of water. Communities have lived through drought, restrictions, floods and supply shocks. Public support for new infrastructure will depend in part on whether water use is responsible, efficient and clearly communicated.

From data to intelligence

Traditional water planning, including long-term demand forecasting, asset condition modelling and supply planning, remains essential. However, data centres and other high-intensity users are increasing the need for a more dynamic view of network conditions.

Utilities need visibility of granular consumption patterns, peak conditions and emerging demand, alongside the ability to integrate datasets such as metering, asset condition, climate, leakage and growth forecasts. This is the practical role of digital water: moving from data collection to operational intelligence.

Closing this visibility gap requires tools that move beyond periodic reads to continuous, high frequency network intelligence.

For high-consumption customers such as data centres, more frequent meter data and predictive tools can help identify demand patterns, detect pressure on local assets and support proactive investment decisions.

Advanced metering and connected sensors can also help utilities capture more frequent network information. AMI with remotely enabled endpoints can transmit data as frequently as every 15 minutes, helping operators identify water loss, monitor consumption patterns and detect potential meter tampering.

That intelligence is most valuable when applied early. The strongest data centre projects consider location, water source, efficiency, reuse, discharge and resilience together, alongside energy, land and connectivity requirements. Recycled water, closed-loop systems and circular approaches may reduce pressure on potable supplies in some regions; however, suitability varies significantly by climate and long-term water availability.

That is why water needs to be part of the earliest feasibility conversations.

A smarter path for digital growth

The rise of AI data centres will test Australia's water planning systems; however, it also gives the sector an opportunity to lead.

Water utilities already understand scarcity, resilience, reliability and stewardship. The next step is ensuring they have the visibility, data and planning influence needed to respond confidently to emerging large-load customers.

That will require earlier engagement, clearer demand disclosure, better cross-sector coordination and stronger use of data-driven tools to anticipate pressure before it builds.

Australia can build the digital infrastructure it needs, provided it plans resource trade-offs transparently before infrastructure decisions are locked in.

The data centre boom should be treated as a catalyst for smarter water management, where every litre is better understood, better accounted for, and used with greater purpose, and where the water sector is at the table from the start.

Nick Phillips is Head of Technical Sales, Asia Pacific at Itron.