Australia’s Data Centre Boom – Love it or Hate it

Australia’s data centre boom is an infrastructure story

The Physical Footprint of Digital Growth

Data Centers have been around for a long time. My first job is 1970 was in the NCR Data Center in Ultimo. It was a non-descript building in Harris Street and operated 24/7, every day of the year. It was noisy, cold and busy. Many larger companies had their own “Data Centers” or Computer Rooms in the day. These were often on show as a corporate statement of size and automation. That soon changed. Today there are data centers that are much bigger and very industrial. Often recognised by the lack of windows, massive air conditioning systems, high security and lots of network connections.

Australia’s data centre boom is often described as a technology story. But it is really an infrastructure story that is driven by technology.

Data centers are becoming the factories of the digital economy. Like factories, they create value. Like factories, they also consume resources.

Behind cloud computing, online banking, government services, streaming, cybersecurity, e-commerce and artificial intelligence sits a growing network of large, power-hungry, water-sensitive and land-intensive facilities. These buildings may look like warehouses, but they are becoming some of the most important pieces of infrastructure in the digital economy.

By 2026, data centres are no longer just hidden server rooms or back-office technology assets. They are becoming part of Australia’s national infrastructure debate. The questions are no longer only about speed, storage and security. They are also about electricity, water, land, local amenity, carbon emissions, cyber resilience, physical security and whether communities are being asked to carry the costs of infrastructure that mainly benefits governments, corporations and the broader digital economy.

Data Centres Australia and DC Byte estimate that Australia had 162 operational data centres as at 31 March 2026, with a further 90 data centres in the pipeline. National operational capacity is forecast to grow from around 1.4-1.5 GW today to about 3.2 GW by 2030. Growth is concentrated in New South Wales and Victoria, particularly around Sydney and Melbourne, where hyperscale cloud and AI demand is strongest. [1]

NOTE: GW = Gigawatt - A gigawatt is 1 billion watts. MW = Megawatt – A megawatt is 1 million watts, TW = 1,000 GW.

3.2 GW is roughly equivalent to the total, continuous electricity required to run more than 2.5 million average Australian homes simultaneously.

The most important point is not only the number of centres. It is their size. The sector is shifting toward fewer, larger and higher-density facilities, including projects above 100 MW. That scale changes the public policy question. It moves data centres from being a private technology asset to being a major user of shared infrastructure.

Indicator	Current estimate or forecast
Operational Australian data centres	162 operational data centres, with 90 in the pipeline
Operational capacity outlook	About 1.4-1.5 GW today, forecast to reach about 3.2 GW by 2030
AEMO/NEM electricity demand outlook	Data centre total demand expected to rise from about 4 TWh in FY2025 to about 12 TWh by 2029-30
Water demand outlook	Industry estimates suggest Australian data centre water use could rise from 5.5 GL (GL – Gigalitre - Billion Litres) to 17 GL over five years (1 GL is the equivalent of 400 Olympic swimming pools full of water)
Global AI/data centre power outlook	IEA projects global data centre electricity consumption could more than double to about 945 TWh by 2030

What is driving the growth?

The first driver is the continuing move from company-owned servers to cloud platforms. Businesses, universities, hospitals, financial institutions, governments and software providers increasingly depend on data centres for storage, processing, backup, security and resilience.

The second driver is digital sovereignty. More organisations want Australian-based storage and processing for privacy, regulatory, latency and national security reasons. This is especially relevant for government, defence-related suppliers, health, finance, critical infrastructure and organisations holding sensitive personal data.

The third and fastest-growing driver is artificial intelligence. AI is often talked about as if it is an invisible service, but it depends on very physical infrastructure: GPUs, servers, cooling systems, substations, batteries, backup generators, fibre links, land and water.

The International Energy Agency projects that global data centre electricity consumption could more than double to around 945 TWh by 2030, with AI the most important driver of that growth. [2]

For Australia, AI changes the scale of the issue. Traditional data centres already require large amounts of energy and cooling. AI data centres require denser computing, more powerful chips, more heat removal and larger, more reliable power connections. That is why the data centre debate is now inseparable from the energy transition debate.

Australia is already data centre dense by population

United States out-ranks all in total number of data centres, but Australia is already high on a per-person (population) basis. Statista’s April 2026 country count placed the United States first with 4,184 data centres, followed by the United Kingdom, Germany, China and France. Australia was listed at 272. [3]

Using Australia’s 2026 population of about 27.2 million, that places Australia at roughly 10 data centres per million people, depending on the dataset used. On this simple population-adjusted measure, Australia sits close to the United States and ahead of larger economies such as the United Kingdom, Germany, France, Japan and China. [3] [4]

This does not mean Australia has a larger data centre sector than those economies. It does not. It means a smaller population base can feel a larger local infrastructure impact. The pressure is not measured only by national totals. It is measured by where facilities are clustered, how large they are, what power and water they require, and how close they sit to homes, roads, schools, hospitals and future housing precincts.

Country	Data centres	Population estimate	Data centres per million people
United States	4,184	349.0m	12.0
Australia	272	27.2m	10.0
United Kingdom	515	69.9m	7.4
Germany	514	83.6m	6.1
France	345	66.7m	5.2
Japan	257	122.4m	2.1
China	369	1,412.9m	0.3

The resource problem: power, water and land

The clearest pressure point is electricity. AEMO’s data centre demand work and related analysis indicate that data centres across the National Electricity Market consumed about 4 TWh of electricity in FY2025, or around 2.2% of grid demand. Under the Step Change scenario, this is expected to rise to around 12 TWh by 2029-30, equivalent to about 6% of NEM electricity. [5] [6]

That creates a difficult timing problem. Data centres can often be planned and built faster than new transmission, renewable generation, firming capacity and distribution upgrades. Australia is already trying to retire ageing coal generation, electrify transport, decarbonise industry and keep power prices under control. Large new 24/7 electricity loads add another layer of pressure.

Water is the second major issue. Data centres need cooling, and cooling can require substantial water depending on the design. Some facilities use more air cooling, some use evaporative cooling, some use chilled water, and some are moving toward liquid cooling and closed-loop systems. The problem is not simply that data centres use water. It is where they use it, whether the water is potable, how transparent the reporting is, and whether demand competes with households, agriculture, ecosystems and housing growth during drought.

The Climate Council cites industry estimates suggesting Australian data centre water demand could more than triple from 5.5 GL to 17 GL over the next five years, with Sydney’s data centre water demand projected to reach 1.9% of city supply and Melbourne’s 0.9% by 2030. [6]

Land is the third pressure point. Large data centres need suitable industrial land, fibre access, major grid connections, water infrastructure, roads, security setbacks and space for backup power. In cities such as Sydney and Melbourne, that land is often close to existing homes, schools, parks, transport corridors or future housing precincts. This creates a planning conflict: should scarce urban land be used for housing, employment, public services or data centres that may create relatively few long-term local jobs compared with their resource footprint?

The issue is not whether data centres are useful. It is whether the infrastructure burden is being planned, priced and shared fairly.

The urban problem: when digital infrastructure lands beside homes

There are good reasons why data centres cluster in and around cities. They need fibre networks, skilled workers, strong power connections, enterprise customers and low-latency links to users. For some workloads, proximity matters. Financial services, critical systems, cloud connectivity and real-time AI inference can all benefit from being close to major population and business centres.

But the disadvantages are becoming harder to ignore. Urban data centres can consume large amounts of electricity in areas where the grid is already under pressure. They can require potable water in cities also trying to support population growth. They can occupy land that might otherwise support housing or higher-employment uses. They can also create local concerns around noise, heat, visual bulk, diesel backup generators, traffic during construction, fire risk and air quality.

Several Sydney councils have raised concerns through the NSW data centre inquiry about blackouts, housing impacts, local health, water demand, noise and cumulative infrastructure pressure. Reporting on the inquiry noted concerns from councils including Ryde, Lane Cove and Penrith about data centres competing with housing, placing strain on water and power infrastructure, and affecting local amenity. [7]

This is where the social licence issue becomes real. The benefits of data centres are often national or corporate: better cloud services, AI capability, sovereign data storage, construction activity, technical operations and broader economic activity. But the costs are often local: water demand, power infrastructure, land use, noise, heat and community disruption.

To look at it from a local, and personal perspective, this area sits in the Lane Cove River bushland precinct which contains a national park with endangered flora and fauna.

That mismatch is likely to become one of the biggest barriers to future growth. The better model may be a layered approach. Some data centres should remain close to cities for latency, resilience and network reasons. But not every AI workload needs to sit beside residential areas. AI training, batch processing and some backup workloads may be better located near renewable energy zones, stronger grid capacity, recycled water availability or industrial areas where heat, noise and land impacts are easier to manage.

Environmental and ecological impact

The ecological impact of data centres is broader than electricity and water. Data centres can affect the environment through carbon emissions if their power is fossil-based, water consumption, heat discharge, land clearing, embedded carbon in concrete and steel, diesel backup systems, battery supply chains, e-waste and pressure on local infrastructure.

The Australian Government has recognised that this is now a national policy issue. In March 2026 it released expectations for data centres and AI infrastructure developers, stating that continued expansion must be environmentally and socially sustainable. The expectations say developers should support Australia’s national interest, minimise adverse impacts on local communities, protect sensitive data, support the energy transition and avoid placing upward pressure on energy prices. [8]

The government also expects data centre and AI infrastructure operators to underwrite new renewable power supply, pay their full share of new grid connectivity so costs are not passed to consumers or businesses, and support the energy transition through demand flexibility mechanisms. [9]

This points to the likely future of approvals. Data centre growth will not stop, but it will become more conditional. Projects that cannot demonstrate additional clean energy, efficient cooling, low potable water use, transparent reporting, responsible land use and proper community engagement are likely to face stronger resistance.

The practical options include recycled and non-potable water, closed-loop cooling, advanced liquid cooling, heat reuse, on-site batteries, firmed renewable power purchase agreements, demand-flexible computing, mandatory water-use reporting, better e-waste management, and planning rules that steer the largest facilities away from sensitive residential interfaces.

How quantum computing may change data centres

Will new technology solve the problem of data centre expansion? Well don't hold your breath. Quantum computing is unlikely to replace conventional data centres in the next five years. It is more likely to become a specialised layer of computing, used for complex optimisation, modelling, cryptography-related research, materials science, chemistry, energy systems and high-value simulations.

If quantum computing becomes commercially useful, it may change the structure of some data centres. Instead of standard server halls only, quantum facilities may need cryogenic cooling, vibration isolation, electromagnetic shielding, precision control systems and specialised technical environments. The resource impact could be positive if quantum computing helps solve difficult optimisation problems, including grid management, battery design, logistics and cooling efficiency. But it could also be negative if large-scale quantum systems require high levels of specialised cooling and energy.

Quantum also changes the security conversation. The Australian Signals Directorate has warned that organisations need to prepare for post-quantum cryptography because a future cryptographically relevant quantum computer would threaten widely used public-key cryptography. For data centres, this means quantum risk is not only about hosting quantum machines. It is also about protecting today’s stored data against future decryption risks. [10]

Cyber and physical security risks

Data centres concentrate enormous amounts of digital value in physical places. A cyberattack on a major data centre, cloud platform or identity system can affect many organisations at once. A physical attack, fire, sabotage event, power failure or fuel incident can also have consequences beyond the building itself.

Australia already treats data storage and processing as part of the critical infrastructure landscape. The Critical Infrastructure Security Centre states that the Security of Critical Infrastructure Act regulates critical infrastructure assets in the data storage and processing industry. [11]

Locating data centres in cities does not necessarily make them more vulnerable to cyberattack. Cyber risk is mainly about systems, identity, software, networks, suppliers and governance. But city locations can increase the consequences of a physical incident because facilities may sit near homes, schools, roads, rail, hospitals, water assets and other critical infrastructure.

This does not mean data centres should be pushed out of cities entirely. Some must remain close to urban users and network hubs. But large urban facilities should be planned as critical infrastructure, not ordinary warehouses. That means stronger physical security, cyber resilience, emergency planning, backup power controls, fire management, fuel safety, separation from sensitive neighbouring uses and transparent risk assessment.

The next five years

Over the next five years, Australia’s data centre sector is likely to grow strongly, but unevenly. The projects most likely to proceed will be those that secure land, power, water, approvals, capital and customers. Sydney and Melbourne will remain major hubs, but resource constraints may push some future development toward regional or peri-urban locations with better access to renewable energy zones, industrial land, recycled water and grid capacity.

The central question is not whether Australia should have data centres. It must. Modern economies cannot function without them. They support AI, cloud services, cybersecurity, banking, research, government, telecommunications, healthcare and digital sovereignty.

The real question is whether Australia can build the next generation of data centres without shifting hidden costs onto households, water systems, electricity consumers, local communities and ecosystems.

Data centres are the factories of the digital economy. Like factories, they create value. Like factories, they consume resources. And like factories, they need rules, planning, transparency and community consent.

Australia’s opportunity is to stop treating data centres as invisible technology and start planning them as visible infrastructure: essential, powerful and valuable, but only sustainable if designed around energy, water, land, ecology, security and the communities that live beside them.

Authors Note on AI Usage

AI is becoming a useful tool and has assisted with this article. I have fact checked all the information as best as possible with my own research and then also AI using ChatGPT Pro 5.5 Extended Thinking. I have also used AI to produce the graphics which I first designed in PowerPoint and then had ChatGPT enhance. I have also used Google Gemini Pro 3.5 Flash to review and find any errors in the information.

References:

1. Australian Data Centre Forecast Report, Issue 1. Data Centres Australia and DC Byte. April 2026. https://datacentres.org.au/wp-content/uploads/2026/04/2026_Australian-Data-Centre-Forecast-Report-Issue-1-1.pdf

   
   

2. Energy and AI. International Energy Agency. 2025. https://www.iea.org/reports/energy-and-ai

   
   

3. Data centers worldwide by territory 2026. Statista. 14 April 2026. https://www.statista.com/statistics/1228433/data-centers-worldwide-by-country/

   
   

4. Population by Country. Worldometer / United Nations Population Division data presentation. 2026 estimates. https://www.worldometers.info/world-population/population-by-country/

   
   

5. Data Centre Energy Demand. Oxford Economics Australia for AEMO. 2025. https://www.aemo.com.au/-/media/files/stakeholder_consultation/consultations/nem-consultations/2024/2025-iasr-scenarios/final-docs/oxford-economics-australia-data-centre-energy-consumption-report.pdf

   
   

6. Seizing the opportunity to do data centres right. Climate Council. June 2026. https://www.climatecouncil.org.au/what-does-the-data-centre-boom-mean-for-australias-switch-to-renewables/

   
   

7. Sydney councils fear new datacentres could cause blackouts, block housing and affect locals’ health. The Guardian. 4 April 2026. https://www.theguardian.com/environment/2026/apr/04/sydney-councils-fear-new-datacentres-could-cause-blackouts-block-housing-and-affect-locals-health

   
   

8. Expectations of data centres and AI infrastructure developers. Australian Government Department of Industry, Science and Resources. 23 March 2026. https://www.industry.gov.au/publications/expectations-data-centres-and-ai-infrastructure-developers

   
   

9. An Australian approach to AI: Expectations for data centres to deliver for Australians. Minister for Climate Change and Energy / Australian Government. 23 March 2026. https://minister.dcceew.gov.au/bowen/media-releases/joint-media-release-australian-approach-ai-expectations-data-centres-deliver-australians

   
   

10. Planning for post-quantum cryptography. Australian Signals Directorate / Australian Cyber Security Centre. 22 September 2025. https://www.cyber.gov.au/business-government/secure-design/planning-for-post-quantum-cryptography

    
    

11. SOCI Act 2018 for data storage and processing. Critical Infrastructure Security Centre. 5 December 2023. https://www.cisc.gov.au/information-for-your-industry/data-storage-and-processing/legislation-regulation-and-compliance/soci-act-2018

    
    

Source note:

Data centre counts vary by definition. Some sources count operational data centres only; others include colocation, cloud, edge or directory-listed facilities. Capacity, energy demand and water demand are better measures of infrastructure impact than facility counts alone.

Author: John Debrincat

© Shapedlogic 2026