25 Feb AI Is Hungry for Power, and Australia Is About to Feel It
KEY TAKEAWAYS
Artificial intelligence might live in the cloud,
but the cloud runs on very real electricity.
Artificial intelligence can feel almost weightless. Ask a chatbot a question and an answer appears seconds later. Upload a document and it can be analysed almost instantly. Businesses move files into “the cloud” without ever seeing where those files physically go. But the cloud is not floating somewhere above us. It is made up of buildings. Very large buildings.
AI Is Driving New Demand
AI, cloud computing and digital services are increasing the amount of electricity required by data centres, making them a much more significant part of Australia’s future energy demand.
Infrastructure Must Keep Up
Growing demand will require more generation, transmission, storage and network capacity to ensure large electricity users can operate reliably as the system changes.
Flexibility Becomes More Valuable
Businesses should consider how future load growth, peak demand, storage, onsite generation and operational flexibility could affect their energy strategy as electricity demand evolves.
Inside them are rows upon rows of servers, networking equipment, cooling systems, electrical infrastructure and backup systems working continuously to keep our increasingly digital economy running. And all of it needs electricity. A lot of electricity. This is quickly becoming one of the most interesting stories in Australia’s energy market.
Energy planning needs to look further ahead
AI demonstrates how quickly new technologies can change electricity demand while energy infrastructure is built to operate for decades. Businesses should consider future load growth, peak demand, available grid capacity, onsite generation and storage when planning expansion so today’s energy decisions can still support tomorrow’s operations.
AEMO now forecasts electricity consumption
by data centres in the National Electricity Market to increase from approximately 5 terawatt hours in 2025–26 to 34 TWh by 2035–36. Their share of electricity supplied through the grid is expected to rise from about 3 per cent to approximately 13 per cent. Put simply, data centres are moving from being a relatively specialised electricity user to becoming a significant part of Australia’s future demand.
“AI may be changing how organisations work.
Behind the scenes, it may also help change
how Australia builds and uses electricity.”
Artificial intelligence is not solely responsible. Cloud computing, video streaming, online banking, e-commerce, corporate software and digital storage were already driving data-centre expansion. AI has simply pressed the accelerator. Training and operating increasingly sophisticated AI models requires enormous computational capacity. The more organisations adopt AI and cloud-based services, the more physical computing infrastructure sits behind the digital world.
For the technology sector, this creates an obvious challenge: where will all that electricity come from? But there is a second question that matters to businesses outside technology. What happens when some of the world’s largest companies become some of Australia’s largest new electricity consumers?
Data centres often require large amounts of reliable electricity around the clock. They cannot simply switch themselves off because the sun has gone down or wind output has fallen. That creates pressure for new generation, transmission, storage and network connections.
AI Is Changing Energy Planning
It can also accelerate investment. Large electricity users can underpin renewable projects through long-term energy contracts. New transmission infrastructure can unlock generation. Storage projects can help balance increased demand. So the AI energy story is not necessarily a story of scarcity. It is a story of enormous new demand arriving while the electricity system itself is undergoing structural change.
There is also an interesting lesson here for ordinary commercial and industrial businesses. For years, electricity demand across developed economies was relatively predictable. Energy planners could look at population, economic activity, weather and industrial demand and build forecasts accordingly. AI shows how quickly that can change.
New demand can accelerate energy investment
Rising electricity demand does not only create pressure on the system. Large energy users can support renewable projects through long-term contracts, new transmission can unlock additional generation and storage can help balance supply with growing demand. The expansion of AI could therefore influence not only how much electricity Australia uses, but also where new energy infrastructure is built.
A technology that barely registered in electricity planning a decade ago is now significant enough to materially influence AEMO’s demand forecasts. That matters because energy investments are long-lived. Power stations, transmission lines, batteries and industrial facilities are designed to operate for many years. Yet the technologies creating electricity demand can evolve much faster.
The consequence is an energy system that increasingly needs flexibility. Businesses should think about flexibility too. Could some electricity consumption move away from peak periods? Could storage help manage demand? Could onsite generation reduce dependence on the grid? Is there enough capacity for planned expansion? What happens if a facility doubles its electrical load over the next ten years?
These questions are no longer relevant only to technology giants. They increasingly belong in ordinary business planning. AI may be changing how organisations work. Behind the scenes, it may also help change how Australia builds and uses electricity.
Sorry, the comment form is closed at this time.