Caroline Sinn and Mia Rebello
Australia is currently grappling with the AI revolution as political leaders debate the development of national AI standards. Every chat, prompt or image created via AI is powered by data centres; the physical infrastructure designed to house, store, process and distribute the massive amounts of data required to generate intelligence.
Currently, 11,426 data centres occupy vast tracts of land in 179 countries, and demand is predicted to triple by 2030. Australia, in particular, has become an attractive destination for data centre development and investment, with more than 160 facilities already operating. This can be attributed to the fact that Australia is a large landmass, has plentiful access to renewable resources, a stable regulatory system, and proximity to the Asia-Pacific market.
However, impacts on energy and water supplies, the housing crisis and Indigenous land rights mean this infrastructure development is not without cost. This prompted Prime Minister Anthony Albanese to announce the legislating of standards for data centre construction, a move aimed at mitigating these harms and showing that, while AI and data centre innovation should be embraced, a more measured approach is needed.
The economic benefits of AI
Up until recently, the Australian government’s response to data centres has been predominantly focused on leveraging the anticipated benefits of AI-related investment, with it expected to reach $100 billion. A significant portion of data centres’ economic benefits can be traced to their positive impact on productivity across the economy. Depending on the extent to which downstream industries adopt cloud computing, artificial intelligence, and data‑driven technologies, indirect broader productivity effects are expected to arise. Data centre construction has been associated with parallel investment of money into attendant infrastructure, including renewable energy projects. Data centres are increasingly considered fundamental core infrastructure for future economies.
In line with this approach, the Labor government’s 2025 National AI Plan is aimed at capturing the opportunity and spreading the benefits of AI, seeking to foster a friendly environment for inbound AI-related investment and ensure investments reach regional areas. This is evident in how much of the plan is concerned with matters like coordinating data centre approval processes to increase the ease and efficiency of starting up these projects.
The strong desire for Australia to be a hub for AI innovation and to benefit from data centre investment is bipartisan. Liberal MP Andrew Hastie recently likened AI infrastructure development to the nuclear arms race of the Cold War era, arguing Australia needs to dramatically scale up investment to retain its strategic independence.
However, experts are divided on the true extent of these benefits. Michael Vardon at the Australian National University argues that while the economic upside is real, it is easily overstated. Data centres are capital intensive, require expensive imported hardware and employ few people, whilst the industries that use AI (often offshore) reap the big productivity dividends. Greater acknowledgement of this fact has arguably led to a shift, particularly from the Albanese government, towards a more measured approach of embracing the innovative benefits of data centres without ignoring and remaining powerless to its impact on the environment and everyday Australians. As the CEO of Data Centres Australia noted, the extent of benefits drawn from data centre construction depends on how the administration outlines its expectations of developers.
Environmental damage from AI usage and data centre construction
Just one widely-used AI chatbot is estimated to process around 2.5 billion prompts per day, consuming hundreds of gigawatt-hours of electricity each year. The US Department of Energy predicts that data centres will consume as much as 12% of all US electricity by 2028. In Ireland, data centres account for more than 20% of total national electricity consumption. Comparatively, Australian data centres consume 2% of electricity in our main grid, but this is predicted to triple by 2030.
Data centres contribute to air pollution by burning fossil fuels for power and creating electronic waste – up to 5 million tonnes will be generated by the expansion of data centres and AI capacities by 2030. To generate the energy required for data centres to run, large amounts of greenhouse gases are released into the atmosphere, directly contributing to global warming and climate change. Just the training phase of one AI model emitted approximately 500 metric tons of carbon dioxide, equivalent to driving a car from New York to San Francisco 438 times.
Data centre cooling systems also run continuously, generating significant noise pollution. In one US study, residents reported between 40-59 decibels of continuous humming noise, impacting their sleep, exercise and concentration. This disrupts communities and can lead to displacement should the noise become unbearable. A Mississippi class action over inescapable noise generated by AI data centre power plants has seen plaintiffs argue the AI boom is wreaking havoc, subjecting thousands of residents to near-constant noise and vibrations.
Impact on housing and water supplies
Companies wanting to build data centres are willing to pay a premium for the opportunity. Planning often directly impacts communities already experiencing housing crises. It exacerbates competition for land, locking out skilled labourers and construction workers and reducing community capacity to meet future housing demand.
In addition, data centres consume anywhere up to 5 million gallons of water per day for cooling, equivalent to the water use of a town populated by 10,000 to 50,000 people. A University of California study demonstrated that roughly one bottle of water is used for each 100-word AI prompt. When AI data centres tap into freshwater resources, they can impact nearby communities.
In Sydney, the 89 existing data centres (at the time of publishing) all draw from Sydney Water’s public drinking supply. Industry and governments are looking into alternatives to consuming potable water for data centre cooling, investigating whether wastewater supplies are a viable alternative.
Belinda Dennett, CEO of Data Centres Australia, told the ABC that if the government mandated the use of recycled water, the industry would be cooperative and even prepared to fund the treatment plants and piping required. However, even if recycled water is used, it is no longer going back into the base flow of rivers and streams, which has ecological impacts and supply issues. Evaporative cooling also leaves behind high concentrations of salts and other contaminants, which can lead to water quality issues.
The UN’s report on the Environmental Cost of Artificial Intelligence indicates that regions facing water or land stresses might bear the brunt of environmental burdens brought about by AI, as communities who live closest to data centre sites may not necessarily be the ones who use or benefit most from AI. In Querétaro, Mexico, expanding digital infrastructure drew on water supplies amid prolonged droughts. With Australia being one of the driest continents on Earth, this concern should be noted.
The construction of data centres may also disrupt private water wells of nearby communities. Residents living next to a Meta-owned data centre in Mansfield, Georgia, complained of an excessive build-up of sediment and cessation of water flow to their houses. Even after plumbers intervened to restore water flow, a brown residue remained.
Impacts on Indigenous rights
Data centre impacts have already been felt in Australia, where existing data centres occupy hundreds of acres of land, rendering it unavailable for farming, recreation and housing.
In parts of Australia, the current speed of data centre development poses serious risks to Indigenous rights to land and resource governance. The data centres and water resources required to run AI systems are often located on or near Indigenous homelands, which raises important questions about sovereignty, environmental impact, and the distribution of benefits and harms. Indigenous communities must be allowed to be involved in decisions affecting their cultural land and resources.
In Boorloo, Western Australia, plans for a data centre were scrapped after resistance from local Aboriginal leaders and communities who argued that industrial development would damage the Mandoon Bilya river system, a site of longstanding cultural connection.
Local community objections
A cluster of five data centres planned for Lane Cove West’s industrial precinct saw residents submitting a record 374 objections against just one of the proposals from the data-centre construction firm, Goodman. In developing data centres, local community consultation is crucial. Although Data Centres Australia claims that it engages effectively with local councils through planning approval processes, this is distinct from genuine engagement with communities. A sentiment shared by many community members is that, irrespective of community-based protests, they are locked out of key secret discussions on data centre development.
Outcry against a data centre proposal in the Blue Mountains resulted in the withdrawal of the development application. The proposed area was surrounded by more than 260,000 hectares of World Heritage-listed national park. Through community action, including a widely signed petition and political campaigning, residents argued low-frequency noise impacts generated by the data centre would be a threat to peace.
Litigation as a vehicle to voice community opposition
Growing trends in climate change litigation reflect community concerns regarding the environmental footprint of large data centres and demanding concessions from big tech. A settlement reached in December 2025 in Center for Biological Diversity v. City of Pittsburg included a commitment that the data centre would use recycled water to cool its servers.
Similarly, in Chile, Municipality of Cerrillos (Google Data Center) v. Evaluation Commission of the Metropolitan Region was filed in 2020. Local residents sought judicial review of the approval of Google’s data centre, claiming that the infrastructure’s enormous water use had not been considered sufficiently in the approval process. Ultimately, the court agreed that the project should be halted.
New developments in the Australian government’s approach
The Australian government has signalled an intention to embrace frontier AI investment but within acceptable parameters. This includes legislating a set of Australian Standards for AI, building on the Data Centre Expectations released earlier this year. These standards will include supporting Australia’s energy transition and require data centres to use innovative, efficient and sustainable solutions to minimise water use. They will demand developers ensure the infrastructure can provide as much power to the grid as they use and meet power and water efficiency demands. Prime Minister Anthony Albanese has also noted that data centres will need to create more jobs than just those required for infrastructure construction. The Office of AI has been established within the Department of Prime Minister and Cabinet to accelerate implementation of the Australian Standards on a national level.
There have been calls to take even more decisive steps. The Australian Greens suggest a moratorium on the approval and building of new AI data centres until communities are meaningfully consulted and environmental concerns considered. This approach was recently taken by New York state, which imposed a moratorium on state permits for new AI data centres in July 2026 to allow for sufficient time to develop a policy around the infrastructure which will reduce energy grid risks, land disruption, excessive water supply use and noise pollution.
A key concern is that developers may simply choose to shift building data centres to other countries, leaving Australia with less digital infrastructure and trailing behind in building the digital capacity needed for research, industry, government services and global competitiveness.
As Australia moves to tackle this AI revolution and become an international hub for data centre growth, key questions remain on the appropriate scope and form of regulation and oversight. Any robust regulatory framework must incorporate community input and address the environmental impacts associated with data centre development. Whether Australia supports innovation and economic growth whilst ensuring the social and environmental costs of AI infrastructure are not overlooked remains an open question.
Caroline Sinn and Mia Rebello were interns with the Australian Journal of Human Rights in Term 2, 2026.
