Unlocking a virtuous cycle: overcoming barriers to AI in Australian energy systems
By Jim Bullock, Director of Power Utilities and Energy Transition, Microsoft Asia, Will Hudson, Director of Energy and Sustainability Policy, Microsoft Asia and Liz Fitch, Head of Government Affairs, Microsoft ANZ
Australia’s energy transition is one of the fastest on the planet. Renewables account for nearly half of electricity, and more than 40 percent of households own Consumer Energy Resources (CER) such as rooftop solar panels, home batteries and smart energy devices. This shift is driving a profound impact on the electricity network.
The grid that was designed and built around a small number of large, centralised power stations now must manage an exponential rise in complexity from variable renewables, two-way power flow, and new digital infrastructure. All of this while keeping power safe, reliable, and affordable.
AI is often discussed as a source of growing electricity demand through data centres. Less attention has been paid to AI’s potential to make electricity systems themselves more efficient, flexible and resilient. The relationship runs both ways. Data centres depend on reliable power, while utilities increasingly depend on cloud and AI to operate, manage and optimise the grid. Each is becoming more critical to the other’s success, and to the resilience of essential services and the broader economy.
Building new transmission, generation and storage remains essential, but these projects take time. AI offers an opportunity to make the infrastructure Australia already has work harder, smarter and more dynamically while new assets are being built.
Yet a new report by Mandala, commissioned by Microsoft, finds that AI adoption across Australia’s electricity market remains uneven. It highlights where AI can add value and what is holding wider adoption back.
What AI can add to the grid
AI is already being used across the electricity value chain, though mostly in small, incremental ways. Generators use it to predict equipment failures before they happen and sharpen forecasts of wind and solar output. Networks combine drone, satellite and sensor data to find faults and manage vegetation before it causes outages. Retailers use it to improve customer service and help households understand how they use energy.
The bigger opportunity is using AI across the whole system in real-time. The report points to the International Energy Agency (IEA), which estimates that AI could help unlock around 175 gigawatts of transmission capacity globally. Grid-enhancing technologies such as dynamic line rating and advanced network optimisation, many of which are themselves AI-enabled, could connect a further 450 to 700 gigawatts of new large loads without building new infrastructure. The IEA also indicates that wider use of AI in energy operations could save around US$110 billion (AUD$158 billion) a year by reducing the number of outages and cost of electricity. Used widely, it could also lower electricity demand by 5 to 10 per cent, and may enable greater ability to connect more renewables.
A system built for stability now needs to become more adaptive
Australia’s grid was built for predictable peak loads. Power flowed one way, from large generators to homes and businesses, and the rules were designed around stability.
The grid of the future is decentralised and two-way. Rooftop solar, batteries, electric vehicles and flexible loads are creating a system with far more assets, data and decisions that need to be made in real time. At the same time, new sources of demand, including large-scale digital infrastructure, are introducing more consistent load into the system and reshaping how stability is managed.
Static rules and human-led processes alone will not keep pace, and the incremental steps taken so far will not get the sector to the grid it needs.
Getting there depends on good data and modern platforms, yet much of the sector still runs on aging, legacy technology. Moving to the cloud is the essential first step. As with digital government, AI only delivers at scale once the foundations are right.
Three barriers are slowing progress
The report found no hard regulatory barriers stopping adoption, only practical and cultural ones. Australia’s technology-neutral approach of applying existing laws is widely supported and gives businesses a workable basis for using AI responsibly today. It points to three main soft barriers.
The first is strategic direction. Many organisations recognise AI’s potential but lack a clear roadmap for where it delivers the greatest operational value and how to deploy it responsibly within existing market frameworks. As critical infrastructure providers, utilities follow strict policy frameworks such as SOCI, and have limited appetite for risk. Clearer guidance from governments and market bodies on how AI fits within the existing rules would give the industry more confidence to move.
The second is investment. Electricity networks are regulated monopolies, and the rules for revenue recovery were built for large physical assets, not software. Current regulatory frameworks were designed to incentivise CapEx investment in physical infrastructure. They are less well suited to digital technologies that can deliver efficiency and reliability improvements through software and analytics. Other markets are adapting. For example, the United Kingdom has adopted a TotEx approach, and set up a fund to reduce the risk of investing in AI in energy, and there is room for investment settings here to value software-based solutions alongside traditional infrastructure.
The third is data. Across the system and often within organizations, it sits in silos, held by individual participants with little incentive to share it. AI needs large volumes of good quality, real-time data, and much of it is fragmented or hard to reach. Work is already under way in Australia and overseas to create secure, trusted environments where data can be shared safely, with privacy protected. Building confidence will also require strong governance around cybersecurity, privacy and responsible AI, ensuring new capabilities enhance rather than undermine trust in critical infrastructure.
A shared opportunity
None of these barriers are fixed, and can be addressed when industry, government and regulators align on a shared ambition. The outcome is a more efficient, reliable and lower-cost electricity system that supports both economic growth and the energy transition.
There is also an opportunity to recognise the role of large-scale digital infrastructure in that future. As demand evolves, hyperscale data centres can provide more consistent and predictable load within the system. As the first hyperscaler to sign up to the Australian Government’s data centre expectations, Microsoft is supportive of the government’s objectives to ensure this growth is managed in a way that supports the energy system and broader economy.
Australia’s energy transition is often framed in terms of physical infrastructure like solar farms, wind turbines, batteries, transmission lines and electric vehicles. These remain essential, but the next phase will be shaped just as much by digital capabilities that enable these assets to be orchestrated and optimized as a system. The ability to manage and optimise millions of connected assets in real time is a non-trivial goal, but will be central to maintaining reliability and affordability.
This is where AI can help Australia do more with the system it already has, moving from isolated projects to system-wide capability. Australia has a strong track record in energy reform and a grid that many countries look to. With the right foundations and a shared ambition to use AI well, Australia has an opportunity not only to modernise its own electricity system, but to demonstrate how advanced economies can use AI to accelerate the energy transition. This would ultimately lead to new jobs, more affordable electricity, and a cleaner more reliable grid for all Australians.
Sources:
47% renewables in NEM (Q1 2026): https://minister.dcceew.gov.au/bowen/media-releases/australias-energy-transition-gathers-pace
40% CER adoption: https://www.climatechangeauthority.gov.au/australias-energy-transition-accelerates-toward-cleaner-safer-grid