Australia’s AI data centre boom is no longer constrained by fibre or latency, it’s constrained by electricity. As the Federal Government considers renewable energy requirements for new data centres, a bigger question is emerging: will energy policy determine which states win the next wave of AI investment?

The rapid growth of artificial intelligence, cloud computing and high-performance computing is driving a new wave of data centre investment globally. Australia is well positioned to benefit from this trend, with abundant renewable resources, political stability, strong connectivity and access to Asia Pacific markets.

The Federal Government’s proposed renewable energy requirements for new data centres are not only about emissions reduction. It also reflects growing concerns that large-scale digital infrastructure could significantly increase electricity demand and place additional pressure on an already constrained energy market.

The rise of AI workloads, particularly large-scale AI model training and inference, is expected to require significantly more power-intensive facilities than traditional cloud computing. In Australia, policymakers are increasingly focused on ensuring that this additional demand does not contribute to higher electricity prices for households and businesses.

The policy intent is therefore twofold:

• Ensure new data centre demand is matched by additional clean energy investment; and
• Minimise the cost of supporting rapidly expanding digital infrastructure from flowing through to existing electricity consumers.

The data centre map is changing

Historically, Australia’s data centre market has been heavily concentrated in New South Wales and Victoria.

Sydney and Melbourne became dominant locations because traditional enterprise, financial services and cloud workloads were highly sensitive to latency. Being physically close to customers, businesses and existing network infrastructure was a major competitive advantage.

This created a hub-and-spoke model, where large amounts of capacity were concentrated around established metropolitan markets. However, the rise of artificial intelligence is changing the equation.

Many AI training workloads do not have the same latency constraints as traditional enterprise applications. Training large AI models involves processing enormous datasets over extended periods, meaning the priority is often not being milliseconds closer to users, it is having access to:

• Fast grid connection timelines
• Large parcels of land suited to expansion
• Significant power and water availability
• Access to network connectivity

This is opening the door for a new generation of data centre locations outside traditional markets. The next AI data centre hubs may not necessarily be built next to Australia’s largest cities, but rather built where energy can be secured quickly and reliably.

The emerging race to connect to the grid

This shift is creating a new competitive landscape between states and territories.

South Australia, Queensland, Western Australia and the Northern Territory are increasingly positioning themselves as future locations for large-scale AI infrastructure because they offer something that Sydney and Melbourne increasingly struggle to provide: access to significant energy resources, water and land.

The competition is moving from: “Who has the lowest latency?”

to: “Who can fastest deliver hundreds of megawatts of reliable power at scale?”

• South Australia has positioned itself around renewable generation, storage and its transition toward a high-renewables grid.
• Queensland has significant renewable energy potential, major industrial precincts and large areas suitable for hyperscale developments.
• Western Australia has unique advantages through its mining-scale energy infrastructure, available land and proximity to renewable energy resources.
• The Northern Territory is positioning itself around large-scale energy projects, including opportunities linked to gas, renewables and regional development.

The common denominator is speed: the jurisdictions that can offer certainty on grid connection timelines will have the edge in attracting AI infrastructure investment.

Will the current policy direction reduce energy prices?

The central question for policymakers is whether additional renewable requirements on data centres will achieve the intended outcome of improving affordability while maintaining Australia’s attractiveness as a destination for digital infrastructure investment.

There is a strong argument that simply limiting or conditioning new electricity demand does not necessarily solve the underlying energy affordability challenge.The National Electricity Market (NEM) is currently experiencing periods of oversupply, with significant renewable generation entering the system and lower operational demand at certain times pushing wholesale electricity prices down.

However, these prices do not necessarily reflect the true long-term cost of building the next generation of energy infrastructure. Current conditions are, in many ways, masking the cost of transition.

Developers of new renewable generation, storage and transmission infrastructure need investment signals that reflect future system requirements, not only today’s surplus conditions. If prices remain suppressed for too long, there is a risk that new renewable projects are delayed because the market signals do not support investment.

This creates a difficult policy paradox:

Australia needs more renewable energy to meet targets, but renewable energy developers need stronger demand and investment certainty to justify building it. Large-scale data centres could provide precisely the demand signal that the energy market currently needs, becoming anchor customers that support new renewable generation, storage and grid upgrades rather than just adding to the strain.

The reality is that Australia’s transition to a cleaner electricity system is unlikely to make energy cheaper in the short term, as ageing coal is replaced and the cost of the new system gets recovered. The long-term objective is not necessarily lower electricity prices immediately, it is a more secure, resilient and sustainable energy system.

The question is how those costs are managed and who helps fund the transition.

A risk of unintended consequences

The risk for Australia is that policy designed to protect energy affordability could unintentionally reduce investment appetite.

Global data centre developers are comparing Australia not only against other states and territories, but against markets across Asia-Pacific, North America and Europe.

For investors, the key considerations are:

• Can I secure enough power?
• How quickly can I connect?
• What will energy cost over 20–30 years?
• Is the regulatory environment predictable?

If renewable requirements create uncertainty or materially increase project costs, some developments may move to jurisdictions where energy can be delivered faster.

However, if designed correctly, the policy could achieve the opposite outcome, encouraging data centre operators to become major investors in renewable energy and helping accelerate the infrastructure required for Australia’s future economy.

The difference will come down to implementation.

Queensland and the Northern Territory are sending a different signal

Perhaps the most interesting development isn’t the policy itself, but the response from Queensland and the Northern Territory.

Both jurisdictions have reportedly resisted the Federal Government’s proposed requirement that new data centres be backed entirely by renewable energy, arguing that such a requirement could reduce their ability to compete for major digital infrastructure investment.

This reflects a broader strategic question: should Australia’s competitive advantage be built around mandating renewable-powered data centres from day one, or around creating the conditions to attract investment first and decarbonise over time?

Queensland and the Northern Territory appear to be leaning towards the latter.

Their opposition to the proposed federal requirements may therefore be less about opposing renewable energy and more about preserving flexibility: developers need firm, reliable power on day one, whether that comes from dedicated renewables paired with storage or, where the grid is still evolving, a mix of renewables and firm generation.

However, there is another side to the equation.

Many hyperscale operators have ambitious global sustainability commitments and increasingly seek locations where renewable electricity is readily available at scale. Jurisdictions with clear pathways to abundant low-carbon energy may therefore become more attractive over the long term, particularly as customers and investors place greater emphasis on emissions performance.

It’s a bet either way: by prioritising flexibility, they may attract developers seeking certainty around energy availability and speed to market. But over time, competitiveness will still depend on demonstrating a credible pathway to lower-carbon electricity as corporate sustainability expectations continue to evolve.

Developers will assess the broader investment proposition: access to power, speed of grid connection, land availability, regulatory certainty, long-term energy costs and confidence that the energy system can evolve alongside their operations.

In that context, the emerging divergence between the Federal Government and some states may become one of the defining factors shaping where Australia’s next generation of AI infrastructure is built.

Australia’s next digital infrastructure winners

The future of Australian data centres will likely be determined by a combination of energy strategy, infrastructure delivery and regulatory certainty.

The winners will be jurisdictions that can offer:

• Fast and predictable grid connections;
• Reliable power at hyperscale volumes;
• Competitive long-term energy pricing;
• Large-scale developable land;
• Clear regulatory pathways; and
• A credible pathway toward lower-carbon operations.

Australia has a once-in-a-generation opportunity to become a major AI infrastructure hub in the Asia-Pacific region.

But the competition is changing. The next wave of data centres will not necessarily follow the traditional Sydney and Melbourne hub-and-spoke model, they will follow electricity.

The policy challenge is finding the right balance: ensuring AI infrastructure supports Australia’s energy transition without making Australia less attractive for the very investment that could help fund the next generation of energy infrastructure.

The states that successfully balance the three competing priorities of investment attraction, reliable and affordable energy, and decarbonisation, are likely to emerge as Australia’s next AI infrastructure investment hubs.

How Australia resolves this tension may determine not only where the next generation of data centres are located, but also how successfully the country captures the economic opportunities of the AI era. The race for AI infrastructure is becoming a race to connect to the grid, and the states that move fastest may define where Australia’s digital economy is built for decades to come.