On the forested slopes of the Otway Ranges lies the West Barwon Reservoir, a source of high-quality drinking water for Geelong and surrounding communities. But the volume of this vital resource is now symbolic of a wider issue: the quiet but massive diversion of drinking water to cool servers in Victoria’s burgeoning data centre industry.

The Concerned Waterways Alliance is deeply concerned about the growing water demands of proposed data centres in Victoria, particularly in Melbourne’s west and north. The ABC reports that documents reveal proposed data centres could collectively consume up to 19.6 billion litres of drinking water annually, enough to empty the equivalent of the West Barwon Reservoir each year, or 80% of the Rosslynne Reservoir in the Maribyrnong catchment where many of these data centres will be located.
This is not just a technical issue or a matter of infrastructure capacity. It strikes at the heart of how we value water in this state, who gets to use it, and how committed we are to restoring our depleted and damaged waterways. At a time when our rivers and aquifers remain in long-term allocation deficit, we must ask: is it acceptable to allocate large volumes of high-quality potable water to multinational data centres while the health of our local rivers like the Maribyrnong, Moorabool, Barwon, Werribee, Yarra, Thomson and Latrobe continues to decline due to over-extraction and insufficient environmental flows?
Victoria’s waterways have borne the burden of extraction for water intensive industries for decades. Now we are at a tipping point. In many catchments, promised environmental flows remain undelivered, and river health continues to deteriorate.
The Victorian government has the opportunity and the obligation to correct this deficit and must take a leadership role. Ensuring the long-term recovery of our rivers must take precedence over enabling new industrial-scale water demands, particularly from sectors where viable alternatives exist.
The Power–Water Trade-off and the Developer’s Equation
Data centres are inherently resource intensive. To keep servers operating safely and efficiently 24/7, enormous volumes of heat must be removed and that cooling requires either water or power. At the heart of this issue is a fundamental trade-off: developers can reduce electricity usage by using more water or reduce water usage by relying on more energy-intensive cooling systems.

In Australia, where electricity prices remain relatively high and our commitment to lowering greenhouse gases has imposed restrictions on inefficient energy use, data centre developers are often incentivised to choose evaporative (water-intensive) cooling, which can use 20–30% less electricity than dry cooling systems. In contrast to and despite the plight of overallocated rivers, water to industry remains relatively cheap. In Melbourne, the cost of potable (drinking) water use for industry is around $3.20 per kilolitre, or less through negotiated agreements, making it economically attractive for developers to opt for water-intensive cooling. This cost is markedly under the cost to households.
However, this short-term cost logic comes with long-term environmental consequences. Water-intensive cooling methods result in the evaporation of large quantities of drinking water, water that should be available for households, agriculture, and the restoration of our struggling rivers. It is a great waste of a critical and scarce resource. While this option might lower power bills for a data centre operator, it increases the burden on the public water supply and shifts costs onto the community, both financially and ecologically.
There are viable alternatives including the provision of a climate independent recycled water supply as an alternative to using drinking water. Dry or hybrid cooling systems require more electricity to operate, but they can be designed to use little or no water. Internationally, some companies are investing in closed loop or liquid cooling technologies that dramatically reduce water consumption, particularly in jurisdictions with stricter regulations or higher water prices. However, in the absence of clear policy in Victoria, it seems many data centre developers are still defaulting to water-intensive designs.
The decision of how to cool a data centre is driven by a range of factors, including:
- Electricity vs water pricing
- Capital vs operational costs
- Climate conditions
- Regulatory obligations (or lack thereof)
| Factor | Low Water Cost | High Water Cost | High Energy Cost | Green Energy Incentives |
| Cooling choice | Evaporative | Dry or hybrid | Water cooling favoured | Dry cooling becomes viable |
| Site preference | Growth corridors with good water access | Industrial zones with recycled supply | Regions with grid upgrades | Areas with renewable power |
| Risk tolerance | More reliant on regulatory stability | Invest in reuse or on-site treatment | May drive innovation in cooling | May absorb higher CapEx |
Unfortunately, Victoria’s current lack of binding water efficiency standards or mandates to use recycled water means there is no pressure on developers to internalise the environmental impacts of their water use. In fact, our permissive regulatory environment and relatively low water costs make us an appealing target for high-water-use developments that would face far stricter scrutiny in other jurisdictions.
Regulatory Gaps and the Need for Leadership
At present, Victoria has no formal requirement for Water Usage Effectiveness (WUE) reporting, no water source hierarchy (e.g., recycled before potable), and no caps or assessments for cumulative water impacts across precincts. In contrast, some European countries and U.S. states such as California now mandate dry cooling in water-stressed regions or require public disclosure of water usage by data centres. Victoria by contrast, has no mandatory water source hierarchy, disclosure regime, or environmental caps on data precincts. This regulatory vacuum invites high-impact development without oversight.
Without policy change, Victoria is at risk of becoming a low-regulation, high-impact destination for global digital infrastructure, where the cloud is cheap to build, but costly for our rivers and communities.
Equity, Water Justice, and Climate Resilience
At a broader level, this is about justice for both our waterways and our communities. When data centres gain access to enormous volumes of relatively cheap drinking water harvested from our rivers, households are left to bear the price both in terms of rising bills and in reduced water availability, particularly during droughts, and our highly stressed waterways remain overallocated.
In Greater Western Water’s catchment area central region, where some of these data centres are proposed, households pay $3.64/kilolitre for the initial 440 kilolitres then a stepped charge of $4.16 per kilolitre after that. GWW’s industry charge is only $3.26/kilolitre, and there are no stepped charges to disincentivise inefficient water use.
With the cost of living already high and many families under stress, we must question why public drinking water should be directed toward the cooling of international tech infrastructure, especially when more responsible solutions exist.
In a drying climate the value of potable water will only increase while water from traditional sources will likely decrease. There is the potential this will drive more frequent water restrictions and it is reasonable to ask whether any curb on household water use would also apply these data centres.
We must also consider the principle of restorative justice for waterways. In many parts of Victoria, rivers and aquifers remain well below sustainable thresholds. Environmental flows have been promised but remain underdelivered. Allowing further industrial extraction of water, especially for non-essential and one-pass use, undermines efforts to heal these systems and betrays the communities who have long fought to protect them.
By contrast, recycled and climate-independent water solutions such as those envisioned in Barwon Water’s “Alternative Water Grid” offer a practical path forward. These projects aim to create a regionally integrated grid that reuses water efficiently, reduces pressure on rivers, and provides certainty for both urban and industrial needs such as data centres. If we are serious about climate resilience, all new high-water-use infrastructure must be required to connect to such networks or demonstrate comparable sustainability outcomes.
What Needs to Change
The Concerned Waterways Alliance calls on the Victorian Government to:
- Mandate WUE reporting and public disclosure for all data centres.
- Prohibit the use of potable water for cooling, unless no other options exist.
- Introduce a differentiated water pricing regime for industry and commerce, that recognises the greater social and environmental value of drinking water.
- Require all new data centres to use recycled or climate-independent water sources as a condition of approval.
- Ensure full accounting of cumulative water impacts across data centre precincts.
- Restore environmental flows and make environmental water recovery a baseline condition before new large-scale uses are permitted.
Conclusion
Victoria stands at a crossroads. We can continue to treat our rivers as a source of low-cost industrial input, or we can recognise them as living entities, essential to the wellbeing of people, landscapes, and future generations. The growth of digital infrastructure should not come at the expense of environmental justice or community water security.
In a drying climate, every drop of drinking water matters and our rivers cannot afford to subsidise the cloud.
The time for policy reform is now. Victoria must show leadership by setting clear standards, demanding transparency, and ensuring that the development of our digital future is grounded in sustainability and fairness. Our rivers and the communities that depend on them deserve no less.
Update
The Concerned Waterways Alliance has released a new position paper on hyperscale data centres, confirming that this rapidly emerging sector now represents a material and growing threat to Victoria’s drinking water supplies.
These facilities operate at a scale not previously accounted for in water planning. Individual campuses can require gigalitre-scale water use each year for cooling, meaning that even a small number of developments can place demand on the system comparable to volumes currently being recovered for environmental flows.
At present, there are no enforceable limits on water use efficiency, no consistent public reporting, and no pricing signals strong enough to constrain demand. In this regulatory vacuum, potable drinking water risks becoming the default supply source—further entrenching reliance on desalination and shifting costs onto households.
While the Victorian Government has begun to acknowledge the issue, including signalling expectations around non-potable water use, this remains policy intent rather than enforceable regulation.
Without clear guardrails such as mandatory water efficiency standards, recycled water requirements, and transparent disclosure the expansion of hyperscale data centres risks locking in long-term pressure on both drinking water supplies and already stressed waterways.