Data centres are now fundamental to the digital economy, housing the servers that store and process information. Yet their rapid expansion raises urgent environmental questions, particularly around energy and water usage.
From streaming a movie to sending an email or using artificial intelligence, almost every digital action relies on something most people never see: data centres.
These facilities are the physical backbone of the internet. As our lives become more connected and data-driven, their importance is growing rapidly. But so too are questions about their environmental impact, energy use and how they fit into communities.
Here's what you need to know.
What is a data centre?
A data centre is a large facility that houses computer servers -powerful machines that store, process and distribute data.
When you open a website, back up photos to the cloud, or use an app, your request is handled by servers in a data centre somewhere in the world. These facilities operate around the clock to ensure digital services are always available.
Some data centres are relatively small and serve a specific business or organisation. Others, known as "hyperscale" data centres, are vast operations run by major technology companies, capable of supporting millions of users at once.
Why are data centres so important?
Data centres underpin much of modern life and the global economy.
They support:
- Cloud computing, which allows businesses and individuals to store and access data online
- Digital platforms, including social media, banking and e-commerce
- Artificial intelligence (AI), which requires vast computing power to train and run models
- Critical services, such as healthcare systems, emergency response and government operations.
As digital technologies expand, so does demand for data centres. The rise of AI in particular is accelerating this growth, as advanced systems require enormous computing capacity.
Why do data centres use so much energy?
Running a data centre requires a significant amount of electricity.
There are two main reasons for this:
- Powering servers - the computers inside data centres are constantly processing data. This requires a continuous and reliable electricity supply.
- Cooling systems - servers generate heat when they operate. Without cooling, they would overheat and fail. Cooling systems - ranging from large air-conditioning units to advanced liquid cooling technologies - are therefore critical, but also energy-intensive.
This growing demand is already becoming a major issue. In Australia, data centres consume several terawatt hours of electricity each year, with projections suggesting this could increase significantly over the next decade.
Hyperscale data centres require extraordinarily high levels of reliability and power quality, operating continuously at scale to support critical digital services and AI workloads.
That means the challenge is not simply about securing more electricity - it's about designing energy systems that can provide resilient, firmed and increasingly sustainable power for a new generation of digital infrastructure.
What about water use?
Cooling data centres doesn't just use electricity - it can also require large amounts of water.
Some facilities use water-based cooling systems to absorb and dissipate heat. Depending on the design and location, this can place strain on local water supplies, particularly in areas already facing water scarcity.
Experts warn that this issue will only become more important over time.
Professor Greg Leslie, Director of the UNSW Global Water Institute, said: "Data centres present a very challenging planning problem for Australia's urban water utilities.
"Unlike other industrial water users, data centres co-locate in clusters to maximise redundancy and have different water demands depending on both the amount of heat generated by the configuration of the servers and the type of cooling system deployed to remove the heat from the chips and the racks.
"Also, as most data centres are designed for expansion, the water supply infrastructure must be designed for ultimate capacity, which may or may not be required depending on the technology used.
"Alternative water supplies, such as recycled wastewater, have an important role to play in reducing the water footprint of data centres, which combined with water efficient cooling technologies and water sensitive precinct design can help support industry growth while reducing pressure on drinking water supplies.
"The Commonwealth's proposed regulations for data centres are a welcome and necessary first step in ensuring efficient use of water resources. The proposal establishes clear expectations around minimising water use, prioritising efficient cooling technologies, ensuring the use of recycled water, meeting water efficiency standards, building resilience to drought and climate change, and transparently reporting water consumption.
"However, the success of these measures will ultimately depend on state and local governments embedding these requirements into planning approvals and operational oversight. It is at that level that we must ensure data centres pay their fair share of infrastructure costs and that the risks, costs and impacts are not shifted onto communities, households, industry or the environment."
Where are data centres located?
Location matters for data centres, and different types have different requirements.
Urban data centres
Some facilities need to be close to cities to reduce delays (known as latency). These support real-time services like video calls, online gaming and financial transactions.
Regional or remote data centres
Larger, hyperscale facilities can often be located further away. These sites may be chosen for access to:
- large amounts of land
- renewable energy resources
- existing industrial infrastructure.
Not all data centres are the same - and this distinction is important.
Gigawatt-scale facilities located in regional clean-energy hubs where industrial-scale renewable energy is more possible may create unique opportunities.
What are the environmental concerns?
The rapid growth of data centres has sparked a range of environmental concerns.
Energy demand and emissions
If powered by fossil fuels, data centres can contribute significantly to greenhouse gas emissions. Even when using renewable energy, their large and growing demand can place pressure on electricity systems.
Water consumption
As noted, cooling systems can require substantial water resources, particularly in hotter climates.
Land use and infrastructure
Large facilities require space, as well as connections to electricity networks, water systems and transport infrastructure. This can lead to competition with other land uses, including housing and industry.
Local impacts
Communities near data centres may experience issues such as noise from cooling equipment, visual impacts from large buildings, and increased demand on local utilities.
Are there ways to make data centres more sustainable?
Yes - but it's complex.
A range of approaches are being explored to reduce the environmental impact of data centres:
- Using renewable energy - many operators are investing in solar, wind and other renewable sources to power their facilities.
- Improving energy efficiency - advances in hardware, software and facility design can reduce the amount of energy needed per unit of computing.
- Innovative cooling technologies - liquid cooling, advanced air systems and even immersion cooling (where servers are submerged in special fluids) can improve efficiency and reduce water use.
- Reusing waste heat - data centres generate large amounts of heat, which in some cases can be captured and reused - for example, to heat buildings or support industrial processes.
- Better system integration - there is growing interest in treating data centres as part of broader energy and infrastructure systems, rather than standalone facilities.
The challenges ahead
Despite these innovations, significant challenges remain.
Demand for digital services is increasing rapidly, particularly with the expansion of AI. Meeting this demand while limiting environmental impact is a major challenge.
Data centres intersect with multiple systems - energy, water, transport and land use. Coordinating these effectively requires careful planning and collaboration between governments, industry and communities.
As facilities grow in size and number, communities are increasingly concerned about local impacts.
There are also broader economic considerations.
"If we start to treat data centres as critical national infrastructure, the question shifts from how we connect individual facilities to the grid to how we design integrated digital, energy and industrial systems that support a resilient, productive and internationally competitive economy," said Prof. Leslie, who is also acting CEO of the UNSW Institute for Industrial Decarbonisation.
"There is a real opportunity for AI data services to become a major export industry for Australia, but we need to distinguish that from the digital and energy infrastructure required to support our own economy.
"Australia's domestic demand for hyperscale AI compute infrastructure is modest relative to larger global markets, which means a significant share of current investment is targeting APAC demand. That makes it critical that export-scale facilities are built on new, clean energy and help expand the system, rather than relying on an already stretched grid."
Why does this matter for the future?
Data centres are set to play an even bigger role in the coming decades - and perhaps especially in Australia.
Australia has the potentiall to turn its growing data centre capacity into a significant new source of economic activity, particularly by selling AI computing power to overseas customers.
If Australia expands its capacity to around 3.1-3.2 GW by 2030, it could generate roughly $US4 billion a year in computing revenue, based on current global averages.
However, Australia would not necessarily capture all of this economic benefit, particularly if data centres are owned by foreign companies.
The overall benefit would depend on how much computing infrastructure is Australian-owned and whether the government introduces measures such as a levy to ensure more of the revenue stays in Australia.
And then there is the issue of our sovereign capability.
"Australia's AI ambitions depend on engineering capability," said Romilly Madew AO, CEO of Engineers Australia and Chair of the UNSW Institute for Industrial Decarbonisation's Independent Advisory Council.
"Engineers will be needed to build and power data centres, strengthen the grid, integrate renewables and ensure this infrastructure is secure and resilient. With demand for engineers growing at the same time as tens of thousands approach retirement and a forecast 20% employment growth, building that workforce must be part of Australia's AI strategy from the outset.
"Data centres are critical infrastructure and we are ready to work with government in taking a coordinated approach across skills, planning, and energy, but a focus must be on bolstering the engineering pipeline."
