Australia has the opportunity to create a new frontier of competitive advantage in sustainable digital infrastructure by aligning the rapid growth of data centres with the expansion of renewable energy and grid integration.
This will become increasingly urgent with the growing adoption of generative artificial intelligence (AI), the need to maintain grid stability during the energy transition, and the imperative to ensure electricity remains affordable for Australian households amid rising demand.
In response to our original article, a representative from the data centre sector offered an important industry perspective on the role of data centres in Australia’s digital future.
We would like to add some further considerations for long-term infrastructure planning in the context of the Australian government’s Future Made in Australia strategy.
The objectives of this strategy and the approach we proposed are to achieve net zero emissions in conjunction with the development of a more complex and resilient economy – both critical to the government’s second term agenda for transforming Australia’s lagging productivity performance.
Current versus future energy consumption
While energy consumption by data centres appears to represent a relatively small share of national energy demand according to AEMO estimates, this masks significant regional variations.
For example, the Greater Sydney region hosts approximately 34 per cent of Australia’s data centres according to industry mapping, meaning the proportion of local electricity consumption will be substantially higher than the national average.
This concentration is particularly important for infrastructure planning, as future growth in generative AI use will create localised pressure points on the grid that require both a scaling of renewable infrastructure investment and strategic management of electricity demand and supply. Housing pressures in urban areas and water supply also come into the mix.
Looking forward, current data centre energy consumption proportions are less relevant than the anticipated upward trajectory. Industry forecasts consistently show that data centre electricity demand is on a steep growth path, driven by increasing generative AI “inference” workloads.
If Australia moves to capitalise on this opportunity and attract foreign investment in AI inference-focused data centres, this growth will accelerate further.
Singapore’s experience offers a valuable lesson. In 2019, Singapore imposed a moratorium on new data centres due to concerns about their significant energy consumption, which at that time already accounted for 7 per cent of the country’s total electricity use.
This decision came as the nation faced growing pressure on limited resources including energy, water and land, while also working toward carbon reduction targets.
The moratorium consequently prompted the industry to develop more sustainable approaches, including improved energy efficiency standards, strategic land use planning and grid integration requirements.
This proactive regulatory approach – addressing constraints before they became critical barriers to industry growth – could serve as a model for Australia.
Rather than reacting to changes as they occur, Australia has an opportunity to proactively develop a staged regulatory framework that maximises the undeniable economic benefits while ensuring future grid stability and energy affordability.
AI inference: The primary driver of future energy demand
The data centre industry noted in its article that “most industry analysts believe AI inferencing is where the majority of data centre future demand will come from in Australia”, confirming the observation in our earlier paper on this issue. This is a clear message for Australia to be well prepared.
Unlike “model training”, which occurs intermittently requiring bursts of intensive compute, inference runs continuously and consumes order of magnitude more electricity than average cloud computing tasks (Figure 1). As AI matures, those ongoing inference workloads will be woven into virtually every industry.
The conventional approach to data centre location emphasises proximity to urban centres, but this overlooks opportunities for strategic co-location with energy infrastructure.
Renewable Energy Zones located outside metropolitan areas could host significant AI data centre capacity while still maintaining high performance.
This flexibility allows for optimizing AI data centre placement for both energy transport efficiency and grid stability while serving users across broad geographic regions.
Grid stability and operational flexibility
The article also rightly notes that data centre operators cannot make unilateral decisions about transaction priorities due to customer agreements. We would accept that data centres cannot and should not be expected to compromise critical services but that does not preclude, for example, requirements for optimising cooling systems and utilising energy-efficient hardware.
Nor does it preclude all forms of demand response. Modern data centres can implement tiered operational protocols that maintain full capacity for critical workloads while adjusting facility-level operations.
Google’s approach to grid events, where they implement algorithms that ‘limit non-urgent compute tasks for the duration of the grid event’, demonstrates that flexibility is possible within appropriate parameters.
The frequency and circumstances of data centre grid disconnections have significant system-wide implications. Thoughtfully designed regulatory frameworks that support grid integration can create mutual benefits for both the data centre industry and the broader public.
By aligning incentives around grid stability and scaled renewable energy investment, regulation can provide predictability for industry planning while ensuring system reliability and cost efficiency – making Australia attractive to foreign investors.
As data centres continue to grow as a proportion of energy demand, coordinated integration approaches will become increasingly vital for maintaining the industry’s social licence to operate and expand.
Forward-looking frameworks can reduce overall system costs, accelerate renewable energy adoption, and create competitive advantages for operators who embrace these approaches.
In the long run, enhanced grid integration serves the industry’s interest in sustainability by ensuring that rapid growth remains compatible with the public interest in reliable, affordable and clean energy.
Looking forward
Australia is uniquely positioned to capitalise on the boom in data centre development. Our nation’s renewable resources, political stability, land mass and growing technical expertise provide a foundation for leadership in sustainable digital infrastructure.
Recent developments in energy storage technology markets show promising trends, with notable price reductions in lithium ion phosphate battery cells and industry analysts projecting ongoing cost reductions, with global cell battery prices expected to reach $80 per kilowatt-hour by 2026 a drop of almost 50 per cent from 2023.
The economic case for this coordinated approach extends beyond direct energy considerations. By strategically co-developing data centres and renewable infrastructure, Australia can create regional technology corridors with built-in stabilising mechanisms.
Data centres provide consistent baseload demand that can be partially modulated to accommodate renewable intermittency, while renewable generation and batteries offer energy security against market volatility – creating system resilience that isolated development paths cannot achieve.
By adopting a forward-looking approach that anticipates how generative AI will reshape energy demands, Australia can create a regulatory environment that serves both industry needs and broader energy transition goals.
This isn’t about imposing constraints on a vital industry – it’s about aligning incentives to create a more resilient, sustainable system.
As Industry Minister Tim Ayres noted in his opening address to a major AI Summit today: “Investments in digital infrastructure like energy hungry AI data centres go squarely toward our national advantages, and our national interest”.
The path forward requires deep and focused collaboration between industry, government and researchers – working together to ensure Australia emerges as a global leader in sustainable digital infrastructure while enhancing grid stability and reducing electricity costs for all Australians.
Dr. Ante Prodan is a computer scientist and complex systems researcher with the School of Computer, Data and Mathematical Sciences at Western Sydney University.
Associate Professor Jo-An Occhipinti is an NHMRC Principal Research Fellow and Co-Director of the Mental Wealth Initiative at the University of Sydney’s Brain and Mind Centre.
Professor Roy Green is Special Innovation Adviser at University of Technology Sydney and former Dean of the UTS Business School.
Do you know more? Contact James Riley via Email.