Australia’s pipeline of huge data centre projects has raised questions about how much energy and water these giant server farms will use.
These questions aren’t simple to answer because we don’t have the data.
We can’t check water or energy use for individual facilities or inspect national data centre statistics and no operator is obliged to disclose these figures.
That may change. Last month, the National Cabinet agreed to develop mandatory national standards for large data centres, covering energy, water and land use. The legislation is due early next year.
No detail is available, and little has been said about their lifetime environmental footprints, from land clearing to noise issues to electronic waste to greenhouse gas emissions. These must be counted too.
The fastest way to settle arguments about the environmental impact of the data centre boom is to collect and publish the data.
Data centres have existed for decades. But something has changed, and this essential infrastructure is now at the centre of a major policy debate.
This article is part of The Conversation’s series on data centres – what they are, why we need them, and why they’re suddenly so controversial.
Hungry for power
In 2020, data centres were estimated to use about 1.5% of the world’s electricity. But since 2022, the artificial intelligence boom has led to rapid growth.
Australia’s 165 operating data centres use an estimated 3% of the power from the National Electricity Market, Australia’s largest grid.
By the mid 2030s, the Australian Energy Market Operator predicts this will reach roughly 13%. This would put them on par with aluminium smelters.
It’s not the national figure that matters, but where, when and how much water is taken.
A new data centre drawing 100 megalitres of water annually from a rural catchment may not matter over a wet winter. In a drought it might.
Some centres are proposed for hot, dry areas where water is already a concern, such as Bundey and Tailem Bend in South Australia. Their water use will depend on the cooling technology used.
At most data centres, only the operator has the data to calculate water use. They are not currently required to disclose use. The water utility sees a connection, not what happens inside the fence. Regulators could look, but it seems nobody has asked them to.
Sydney’s data centres currently draw about 3.5 billion litres a year, less than 1% of the city’s water supply. But this could rise to 25% within a decade, according to Sydney Water estimates.
By comparison, industry modelling suggests a figure of 2% by 2030.
This tenfold difference in forecasts for the same city is due to the fact much has to be assumed given the lack of data.
The economic benefits of data centres are largely national or international, but environmental impacts are felt locally. Pictured: protesters against data centres in Melbourne in early September.Mitchell Costello, CC BY-NC-ND
Trade-offs and complexities abound
If a data centre runs on renewables and uses the same water in a closed loop or new direct-to-chip cooling, the impact will be much smaller than using evaporative cooling, where much water is lost.
Similarly, power drawn from fossil-fuel dominated grids means water is used indirectly, as coal and gas plants use lots of water – especially compared to renewables.
The trade-off is direct: less water for cooling usually means more electricity.
While energy efficiency is improving and advanced liquid cooling can cut water use by up to 50%, this isn’t a full solution. Australia could build giant water-efficient data centres and still end up with climbing water use.
Data centres use lots of power to run their servers and require water to keep them cool.GoAerials/Getty
Lessons from abroad
In Ireland, data centres use 23% of electricity, close to all households combined (28%). But data centre demand is rising fast, jumping 10% last year – while household demand rose 2%. We know this because Ireland’s statistical office measures it. Ireland measures. Australia estimates.
Ireland’s transparency lets policymakers make hard decisions such as limiting new grid connections and requiring the use of renewables to meet 80% of demand.
The main lesson from the United States is that the most intense impact is often local. Communities chosen to host new data centres often push back hard based on issues around noise, energy and water use, and visual impact. Backlash against new data centres is becoming intense in many areas – and bridges political divides.
At a national scale, today’s data centres have modest environmental costs and potentially large economic benefits.
But we don’t know what this looks like when translated to local impacts.
The real issue is what’s in the pipeline. Right now, 225 data centres are proposed or under development. If all are approved, this would represent 67 gigawatts of new power demand – a massive jump on the 2.2 gigawatts currently operating.
Some of these projects won’t get up. Over a third of the planned projects listed in 2025 were cancelled. But some will be built. Because company disclosures are voluntary, selective and unaudited, they cannot be compared with each other or checked.
As a result, anxiety is filling the space where data should be.
What should Australia do?
Ideally, new data centres would draw power from new renewables with minimal indirect water use.
At a federal level, authorities want large data centres to source their own clean power. But Queensland and the Northern Territory want to use gas. This issue is yet to be resolved.
For water, maximising efficiency is the mantra. While efficiency is good, it is not enough. Water is extracted, not efficiency. What matters is total use against what is available and how much competition there is.
Energy use is expected to include reporting, a regulator and testing. Water use will need the same, and the standards should include noise, electronic waste and emissions. If, as the industry and government claim, there is no need to worry, neither should object to such standards.
Done poorly, Australia’s data centre boom could suck up scarce water supplies and increase the use of fossil fuels. But if done well, we could grow the industry using clean power and ensuring water use is monitored and minimal.
The difference comes down to having data. Much of the extra data we need to track data centre impact can be gathered by the Australian Bureau of Statistics – just as Ireland has done.
Transparency can help the data centre industry improve its social licence to operate. With data we can see if the largely national and international economic benefits of data centres justify the mostly local environmental costs.
Michael Vardon is affiliated with the Society of Natural Capital Accounting. These views are his own. He has previously worked for the Australian Bureau of Statistics.
Since the end of the second world war, adding phosphorus to soils in the form of mineral fertiliser has played a crucial role in increasing and maintaining global food production in response to population growth.
The unused phosphorus is mainly stabilised through chemical reactions which bind it to soil minerals. It accumulates as “legacy soil phosphorus”, with an increased risk of contributing to water pollution.
In New Zealand, phosphorus for fertilisers is manufactured exclusively from imported phosphate rock. But there is an alternative to using rock-based phosphorus.
Struvite is a phosphorus-rich compound that can be recovered from wastewater treatment processes.
As our research shows, struvite raises phosphorus uptake by plants and may reduce transfer to waterways. It could at least partially replace conventional phosphorus fertilisers.
A field fertilised by struvite.Florencia DeLucca Agrelo/Lincoln University, CC BY-SA
For example, after decades of relative stability, the cost of single superphosphate has doubled in price from NZ$200 to $400 per tonne since 2000, with notable spikes in 2008 ($480 during the Global Financial Crisis), 2022 ($500 at the start of the Russian invasion of Ukraine) and 2026 ($550 because of Middle East conflicts).
It is widely acknowledged that we need to improve the efficiency of phosphorus use in primary production systems. This can be achieved by using different plants in crop and grassland systems to enhance uptake from fertiliser and legacy soil phosphorus, and by recovery and reuse of phosphorus from waste.
Struvite can be recovered at scale from nutrient-rich municipal wastewater and industrial waste from sources such as milk processing. The technology for isolating and purifying struvite from wastewater is well established.
Struvite removal from wastewater also benefits the environment by reducing direct discharge of phosphorus and nitrogen into waterways.
Better uptake by plants, lower loss from soil
Research from North America and Europe shows that struvite can be as effective as conventional mineral fertiliser as a source of phosphorus for plants.
However, most of this research has been carried out on annual crops grown in neutral to alkaline soils (pH higher than 6.5). New Zealand primary production is mainly based on perennial mixed-species grasslands grown in acidic soils (pH lower than 6.5).
We recently completed a comprehensive investigation of struvite as a source of phosphorus for New Zealand grasslands. This involved a combination of field, controlled-environment and laboratory experiments, and included soils with different phosphorus retention capacities.
This means that using struvite instead of single superphosphate, a similar plant yield could be achieved with a third less phosphorus added to the soil.
Our findings are consistent with data from a three-month glasshouse experiment carried out on two acidic grassland soils with contrasting relative phosphorus retention capacities (12% and 96%).
On average, plant phosphorus uptake was 49% higher from struvite compared with single superphosphate, especially for soils with a lower phosphorus retention capacity.
Soil analyses conducted at the conclusion of both the field and glasshouse experiments revealed that concentrations of potentially plant-available phosphorus were higher when struvite was applied.
Further confirmation came from a laboratory incubation experiment carried out on three acidic grassland soils with different phosphorus retention capacities (12%, 45% and 94%). Results show, on average, increases in plant-available phosphorus 63 days after application were markedly higher for struvite compared with single superphosphate.
Finally, results from a controlled-environment study reveal that struvite significantly reduces immediate phosphorus loss from soil through drainage compared with single superphosphate. Total phosphorus measured in leachate from a soil with a low phosphorus retention capacity was 80% higher for applications of single superphosphate compared to struvite.
Overall our findings clearly demonstrate that struvite recovered from wastewater could be a viable and sustainable alternative source of phosphorus fertiliser for primary industries in New Zealand.
The next step is to quantify how different soil types, plants and environmental variables influence the effectiveness of struvite phosphorus and to assess its commercial viability.
Leo M Condron does not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.
Australia’s only nuclear reactor is celebrating 20 years of producing vital ingredients for lifesaving medicine, critical electronic products and groundbreaking research.
To date, the world has failed to tackle climate change head-on. We must now focus on limiting the magnitude and duration of “overshoot” – how much global temperatures exceed the 1.5°C limit. Every fraction of a degree our planet warms, and every year spent there, will inflict irreversible ecological damage.
A new report, developed by my research team at Climate Analytics and the Potsdam Institute for Climate Impact Research, asks a simple question: what can Australia do to limit this overshoot?
Our latest research suggests the world can still limit peak warming to about 1.7°C, reach net zero greenhouse gas emissions in the 2060s, and bring warming back below 1.5°C by 2100. But only if governments take urgent and ambitious action.
Fossil fuels – including oil, coal and gas – account for more than 75% of global greenhouse gas emissions. To curb climate change, we must rapidly transition away from their use. Australia committed to this goal at COP28 in 2023, at COP30 by signing the Belém Declaration and at the Santa Marta conference in April 2026.
In Australia, we’ve had too many reminders of what’s at stake. The horrific Black Summer fires of 2019–20. Repeated record floods, including across Queensland and northern New South Wales following the 2017 Cyclone Debbie. Not to mention the mass bleaching of the Great Barrier, Ningaloo and Scott reefs.
Now, a potentially record-shattering El Niño is forming. Research suggests it could temporarily push global warming towards 1.7°C or even 1.8°C in 2027, likely bringing extreme heat and drought conditions to Australia.
Our new report shows Australia can realistically meet our climate targets. But that demands we urgently adopt an ambitious emissions pathway, built on the Paris Agreement.
Australian firefighters are increasingly having to prepare for severe bushfires.Aston Brown/Getty
The path forward
A Paris-aligned emissions pathway would require Australia to transform its energy system.
Fossil fuels
Fossil fuels supply about 90% of our country’s raw energy and make up a large share of our greenhouse gas emissions. Australia is also one of the world’s biggest fossil fuel exporters, producing roughly 4.5% of global carbon dioxide emissions.
To achieve our suggested pathway, Australia’s net emissions would need to fall 55% below 2005 levels – the year on which Australia’s national targets are based – by 2030 and reach net zero by 2050.
But under current policies, Australia will barely meet its 2030 target of a 43% reduction. Worryingly, this means we will likely arrive at 2050 with 245 million tonnes of residual emissions still to be offset – more than half of today’s net emissions. Australia would also need its coal and liquefied natural gas production to drop by about 80% and 25%, respectively, by 2035. Coal exports would have to fall by 95% before 2040.
Renewables
Decarbonisation – switching from fossil fuels to renewable energy sources such as wind and solar – is critical to staying below the 1.5°C limit.
To achieve this, electricity must make up 50% of Australia’s national energy demand by 2035, up from 25% in 2025. However, current policies would only deliver 33% by 2035, so the 35% by 2035 electrification goal Australia and Turkey are putting forward at this year’s COP31 summit is simply business as usual.
Sheep graze around solar panels at the Dubbo Solar Hub, west of Sydney.Gregory Plesse/Getty
Policy must keep up
Achieving these reductions would require major policy changes.
For starters, the current non-binding target of 82% renewables by 2030 – which Australia is likely to miss according to the Australian Energy Market Operator – must be strengthened and made binding. This would make it possible to speed up the renewables rollout, potentially reaching 96% renewables by 2035.
The Safeguard Mechanism – a federal government policy that sets legally binding pollution limits on industrial facilities – must be properly enforced. In 2025, industrial emissions appeared to fall by just 2.3% – less than half the 4.9% annual decline needed. What’s more, data from the 196 facilities covered by the Safeguard Mechanism in both 2023–24 and 2024–25 shows their combined emissions dropped by just 0.4%. That’s because the mechanism effectively offers companies unlimited offsets that decrease their net emissions.
The moment to fix this is the 2026–27 Safeguard Mechanism review, which is currently receiving stakeholder feedback.
In the oil-dominated transport sector, there are calls for the government to end new petrol and diesel car sales by 2035. However, it has no current plan or policies to electrify heavy vehicle freight. Emissions from mining – one of the few industries where fossil fuel use is still rising – could be pulled down to zero by 2045 if we replace diesel with electricity in trucks, equipment and processing. Scrapping the diesel rebate and other fossil fuel subsidies would also help.
An Australian fossil fuel phase-out would be a signal of genuine leadership, particularly among our Pacific neighbours already grappling with rapidly rising seas.
Yet, the Albanese government has approved 37 new, expanded or extended fossil fuel projects since coming to power. It has 42 more in the pipeline for approval.
This goes directly against the International Court of Justice which, in its 2025 advisory opinion on climate change, found licensing new fossil fuel projects can constitute an “internationally wrongful act”.
Australia is set to lead talks at COP31 in November and has vowed to continue the push for a fossil fuel transition. However, we cannot demand that others respect climate targets we ourselves are failing to meet.
We already have a pathway to keep global warming below 1.5°C. What’s missing is the political will to take it.
The Pacific nation of Tuvalu is vulnerable to sea level rise.Mario Tama/Getty
Bill Hare receives funding from Bloomberg Philanthropies, the IKEA foundation and the European Climate Foundation.
