Powering the Machines: Why Data Centres Could Become the Next Big Energy Trade
Artificial intelligence may grab the headlines, but electricity will write the cheques.
The International Energy Agency’s latest report, Key Questions on Energy and AI, puts hard numbers around something many suspected for some time: the digital economy is becoming a major energy consumer. (1) We knew the data needs were large. We now know they are enormous.
For investors, this matters. Every extra megawatt required by a data centre needs generation, storage, networks, cooling systems, transformers, land and capital. AI may be built on software, but it runs on steel, copper and electrons.
What Is the Report About?
The IEA’s report examines the relationship between artificial intelligence, data centres and global energy demand. It updates prior modelling and looks at how rapidly growing computing needs are reshaping electricity systems, infrastructure planning and investment priorities.
The standout figure from the report is simple enough. Data centre electricity consumption grew by more than 15% between 2024 and 2025, adding around 70 terawatt-hours (TWh), taking annual use to almost 500 TWh, or just over 1.5% of global electricity demand.
That is not a rounding error. That is the electricity equivalent of an industrial sector arriving in a hurry.
The IEA’s Base Case sees data centre electricity consumption doubling by 2030. In more aggressive scenarios, demand rises further.
This growth is being led by hyperscalers, the giant operators behind cloud computing and AI platforms. According to the report, the combined electricity consumption of four major hyperscalers reached 110 TWh in 2024.
Why Is Demand Rising So Quickly?
Traditional internet use was already energy intensive. AI adds another layer.
Training large language models, running image generation, processing video, and managing increasingly sophisticated reasoning systems all require powerful chips operating at scale. These systems generate heat, need constant uptime, and often demand far more power density than previous computing loads.
As per The Energy, AI-focused data centre power consumption increased by 50% as high-density server racks drive a 44-fold surge in power density from 2020 to 2027. (2)
In plain English, the machines are getting hungrier.
Efficiency gains are helping, but not enough. Better chips and smarter software reduce energy per task, yet total demand keeps climbing because usage growth is overwhelming those savings.
A familiar story. Cars became more fuel efficient, then people bought bigger cars and drove further.
Why This Creates Pressure on Energy Systems
Electricity grids were not designed with an army of 24-hour AI warehouses suddenly knocking on the door.
The IEA notes that policy makers and system operators face challenges around grid connections, planning delays, transmission upgrades and affordability impacts.
In some markets, developers are being forced to think creatively.
This includes:
Behind-the-meter battery systems.
Onsite gas generation.
Flexible demand arrangements.
Locating in regions with spare grid capacity.
Long-term renewable power purchase agreements.
As per The Energy response to this report, around 20GW to 25GW of battery energy storage systems could be installed in data centres by 2030.
That is significant in its own right. Data centres are no longer just customers of the grid. They are becoming mini power systems.
The Opportunity for Investors
This is where the story becomes interesting.
The obvious winners may be data centre owners and AI software groups. But the second-order opportunities could be broader and, in some cases, more durable.
Renewable Energy
Large technology companies remain among the biggest buyers of renewable power globally. They need long-term, scalable and increasingly low-cost electricity.
Wind, solar and hybrid renewable projects with storage may find a deep-pocketed customer base.
Energy Storage
If data centres need reliable power and protection from grid constraints, batteries move from nice-to-have to essential.
Short duration batteries for peak management, backup systems and grid support all stand to benefit.
Networks and Grid Infrastructure
Transmission lines are not glamorous. Neither are substations or transformers. They are, however, increasingly valuable.
The IEA specifically notes supply chain risks around transformers and related equipment.
Investors often chase the shiny object. Sometimes the money is made selling picks and shovels.
Cooling, Efficiency and Specialist Equipment
Moving data is one challenge. Cooling the servers doing the work is another.
Efficient cooling systems, power management technology and next-generation semiconductors may all benefit from sustained demand growth.
Real Assets and Private Markets
The IEA also notes that data centre investments have grown too large to be funded from company balance sheets alone.
That suggests room for infrastructure funds, private credit, co-investment vehicles and institutional capital. In short, this buildout may need Wall Street, super funds and pension capital as much as it needs coders.
What It Means for Australia
Australia has an interesting hand to play.
We have abundant renewable resources, growing battery deployment, strong engineering capability and increasing interest in digital infrastructure. Regions with low-cost renewable energy and available land may become more attractive locations for future data centre investment.
If policy settings are sensible, Australia could export more than commodities. It could export computing capacity powered by clean energy.
That would be a useful twist in the national story.
The Bottom Line
The IEA report confirms what markets are only starting to price in: AI is not just a software boom, it is an energy boom.
Data centres are becoming one of the largest new sources of electricity demand in the global economy. Meeting that demand will require vast capital spending across generation, storage, networks and efficiency technologies.
For investors, the opportunity may not only be in the companies creating the algorithms. It may sit with those supplying the power, the hardware and the infrastructure that makes the whole machine run.
The digital future, it turns out, still needs cables in the ground.
References
International Energy Agency, Key Questions on Energy and AI, 16 April 2026. https://www.iea.org/reports/key-questions-on-energy-and-ai
D. Braue, AI data centres risk becoming victims of their own success, IEA warns, The Energy, 19 April 2026. https://theenergy.co/article/ai-data-centres-risk-becoming-victims-of-their-own-success-iea-warns
Important Information
EnviroInvest Pty Ltd ACN 685 107 957 (“EnviroInvest”) is an Authorised Representative of Daylight Financial Group Pty Ltd ACN 633 984 773 (“DFGPL”) which is the holder of an Australian Financial Services Licence (AFS Licence No. 521404).
Information in this commentary is current as at date prepared unless otherwise stated. However, please bear in mind that investments can go up or down in value, and that past performance is not a reliable indicator of future performance. For more Important Information please refer to the Disclaimer section of this website.
This communication may contain general financial product advice. It has been prepared without taking into account your personal circumstances, and you should therefore consider its appropriateness in light of your objectives, financial circumstances and needs before acting on it.
If our advice relates to the acquisition or possible acquisition of a particular financial product, you should obtain a copy of and consider the Information Memorandum (IM) or Product Disclosure Statement (PDS) before making any decision.