Between March and August 2026, the Australian Government rewrote the playbook for how large-scale data centres are regulated.
For the first time, AI infrastructure developers face rules that treat their facilities the way heavy industry has always been treated: as a serious draw on water, land and energy.
In this article, we’ll walk through what the new national framework requires, why water and land have joined power as the central environmental concerns and what it all means for businesses planning, investing in or approving data centre development in Australia.
Fast overview: what just changed for AI data centres in Australia
In March 2026, the federal government published its “Expectations of Data Centres and AI Infrastructure Developers,” setting out voluntary best-practice standards across energy, water and community impact. By July, the conversation had escalated and mandatory standards were flagged for:
- Power usage and generation
- Potable water use
- Local environmental risks
National Cabinet discussions in August 2026 are set to formalise these expectations into a binding National Framework.
Why AI data centres are suddenly an environmental issue in Australia
The rapid expansion of artificial intelligence has turned what was once a niche energy efficiency question into a multi-dimensional environmental risk issue. Generative AI demands extreme compute density and that translates directly into rising energy and water consumption.
The numbers tell the story. Data centres in Australia used 4 TWh of electricity in 2024–25; consumption is projected to triple to 12 TWh by 2030. Australia’s National Electricity Market could see data centres using up to 6% of its electricity by 2030.
Environmental pressures from data centres include carbon emissions and water consumption and they often require backup systems (diesel generators, battery arrays, uninterruptible power supplies) that contribute to embodied emissions and electronic waste issues.
Energy, water and land: the three pillars of the new Framework
The National Framework for large-scale data centres is expected to rest on three environmental pillars. Each one affects how developers plan, build and operate.
- Energy: Data centres will be required to generate or underwrite equivalent renewable energy to offset their electricity use. Data centre energy efficiency is measured using NABERS ratings and Power Usage Effectiveness (PUE) and these benchmarks are likely to become part of mandatory reporting. The Australian government targets 82% renewable electricity generation by 2030, and large data centre operators will need to demonstrate alignment with that trajectory.
- Water: There will be strict performance benchmarks on potable water use for cooling, along with a strong preference for recycled water or alternative cooling technologies. Metering and transparent public reporting of water consumption will be required.
- Land: Expectations around site selection will require developers to avoid sensitive environments and to integrate with local planning schemes and community expectations.
The government has signalled a preference for clustering large-scale AI data centres where renewable energy, recycled water and suitable industrial land already exist, rather than allowing ad hoc urban fringe development.
More detailed technical standards, potentially including AI-specific metrics beyond current NABERS for data centres, are expected to be developed as part of implementation.
Water use rules: what “minimising water” will mean in practice
The most immediate environmental change for AI data centres is around cooling water. Data centres require significant volumes of water for cooling, impacting local water resources.
To put the numbers in context: data centres currently use less than 0.1% of Australia’s total water. But data centre water demand may triple from 5.5 GL to 17 GL in five years, with demand projected to triple to approximately 17 GL by 2028. In Sydney, that means data centres will use 1.9% of the city’s water supply by 2030.
Environmental due diligence will increasingly need to assess impacts on local groundwater as well as risks of contaminant migration from cooling water discharges, leaks and chemical storage.
Land and groundwater risks specific to Australian data centre sites
While most AI infrastructure discussions focus on electricity, the most complex environmental risks are often under the surface: soil conditions, groundwater behaviour and legacy contamination.
Key Australian land-based risks include:
- Former industrial land or filled sites with unknown site contamination that can create environmental liability if not identified early.
- Proximity to wetlands, rivers, estuaries or coastal lowlands where acid sulfate soils and high groundwater tables are common. This is especially relevant to the Perth metro area and eastern seaboard floodplains like the Wianamatta-South Creek system in Western Sydney.
High loads and 24/7 operation make data centre infrastructure particularly sensitive to flooding and stormwater inundation, soil movement and settlement and corrosion from aggressive groundwater conditions such as acidic or saline water. Regulators will increasingly expect a full environmental risk assessment covering these issues before approving large data centre proposals, rather than treating them as low-impact commercial buildings.
Environmental due diligence: data centres treated like heavy industry
In short, the new national rules effectively move large-scale data centres into the same risk category as industrial plants when it comes to environmental due diligence.
Environmental due diligence in this context means pre-acquisition and pre-development investigations to identify contamination, soil and groundwater constraints, flooding and other environmental risks. It also means assessing compliance obligations under state contaminated sites legislation, planning schemes and environmental protection laws.
Assessments typically follow a two-phase process:
- Phase 1 includes a desktop review and site inspection: historical land use, aerial imagery, regulator records and identification of nearby sensitive receptors.
- Phase 2 involves soil and groundwater sampling and testing, with laboratory analysis to detect contaminants and assess geotechnical and hydrogeological conditions.
Environmental assessments help manage contamination liabilities before transactions. Lenders, investors and foreign investment review processes are increasingly requiring this level of environmental site investigations and due diligence for AI infrastructure, particularly where substantial water and power use is involved.
What a comprehensive site investigation for an AI data centre should cover
The scale and sensitivity of AI infrastructure means standard commercial property checks are no longer enough. A purpose-built site investigation is needed and environmental site assessments are often required for compliance.
Key technical components include:
- Soil profiling to identify contamination, geotechnical constraints and potential acid sulfate soils.
- Groundwater level monitoring and quality testing to understand seasonal fluctuations, potential contamination, salinity, pH and interactions with proposed foundations, service trenches and cooling water systems.
- Historical and off-site contamination source assessment: mapping former fuel depots, industrial estates, landfills and other potential sources of contaminant migration onto the property. Additional investigation may be warranted where off-site plumes are suspected.
- Surface water and drainage assessment, including flood risk under current and future climate change scenarios.
- Evaluation of proposed fuel storage, chemical handling and waste management to prevent future contamination from diesel, cooling chemicals and firewater runoff.
How environmental risk and due diligence shape approvals
Environmental risk is no longer a box-ticking exercise for data centre operators. It is a material factor in planning approval, grid connection and community acceptance.
Environmental risk assessment now interacts directly with:
- State and local planning approval processes, including rezoning, development approvals, and conditions related to groundwater, stormwater, noise and visual impact.
- Federal and state environmental legislation where matters of national environmental significance or significant contamination issues arise.
- Assessment by water utilities and electricity networks of potential impacts on local infrastructure and climate targets.
Poorly understood environmental risks can delay or derail projects entirely, triggering legal action, cost blowouts and loss of community trust. Raising awareness of these risks early is far cheaper than managing the adverse effects later.
Get in touch
If you’re concerned about data centres’ environmental impact in Australia or contamination of any kind on or around your property, our consultants are here to help.