
Power first: why Australia's data centre boom is a capital and sustainability decision — not a property play
Australia's data centre boom is a power and capital problem, not a property play. Demand could hit 12 TWh by 2030, yet grids are constrained and 83% of new global demand may be met by fossil fuels. Why energy strategy is now project value.
Australian data centre electricity demand is forecast to grow around 25% a year to roughly 12 TWh by 2030 — about 6% of the National Electricity Market — with AI-optimised servers alone set to exceed a third of local data centre consumption in 2026 (AEMO/Oxford Economics; Gartner). Yet the grid is constrained, connection timelines run to two years or more, and Bloomberg NEF estimates around 83% of incremental global data centre demand could be met by fossil fuels by 2030 without deliberate intervention. For developers, operators and investors, energy strategy has become the single largest determinant of both project value and sustainability outcome. It is a capital decision, and it belongs in the finance function.
Data centres used to be valued like real estate — servers, cooling, fibre, security. They are now underwritten like energy-infrastructure projects. In Australia, whoever secures firmed, low-carbon power on defensible commercial terms will decide who builds, who returns capital, and who earns a social licence.
The demand surge is real — and concentrated
The numbers are large and moving fast. AEMO, using Oxford Economics Australia modelling, forecasts data centre consumption growing at about 25% a year to 12.0 TWh by FY30 — roughly 6% of the NEM — and 34.5 TWh by FY50. Gartner puts Australian data centre electricity at 4.5 TWh in 2025, rising 38% to 6.2 TWh in 2026 and reaching 15.7 TWh by 2030, with AI-optimised servers set to exceed a third of local consumption in 2026. Australia already hosts more than 250 data centres, and national occupancy rose roughly forty-fold between 2005 and 2025, two-thirds of it since 2020 (United States Studies Centre).
The demand is also highly concentrated. In New South Wales, data centres account for around 4% of grid-supplied electricity today, rising toward 11% by 2030; in Victoria, 2% rising toward 8%. In March 2026 the NSW Investment Delivery Authority prioritised 15 data centre projects worth A$51.9 billion, while AusNet reported assessing more than 10 GW of connection requests in Victoria alone.
The real constraint is power, not property
The binding constraint is no longer land or capital — it is firmed, connected power. AEMO reported 11 large projects representing 5.4 GW of maximum demand moving through the transmission connection process at the end of the March 2026 quarter, with connections typically targeting a two-year timeline from application to energisation, and facilities ramping over five to ten years.
The pipeline also contains a large measure of what the Climate Council calls "phantom demand." In NSW, 44 data centres representing 11.4 GW sit in the pipeline — equivalent to nearly four Eraring power stations — yet industry expects only about 1.2 GW to actually come online in Sydney by 2030. Speculative connection applications inflate the apparent pipeline and complicate planning. In this environment, a credible, secured power position is what separates a project that closes from one that stalls.
The sustainability reckoning has arrived
The sustainability exposure is now both financial and political. Because data centres run flat, continuous loads, their demand profile favours baseload generation — which is why BNEF estimates around 83% of incremental global data centre demand could be met by fossil fuels by 2030 absent deliberate procurement. In Australia, where renewables supplied about 40% of generation in 2024, a data centre drawing from the prevailing grid mix inherits a materially carbon-intensive Scope 2 profile.
That profile now carries consequences. Under AASB S2, purchased-electricity emissions are a Scope 2 disclosure for operators and, through leased and value-chain arrangements, a Scope 3 exposure for landlords, tenants and financiers. On 23 March 2026 the Australian Government released its Expectations of data centres and AI infrastructure developers, signalling that proposals aligned with energy, water and national-interest expectations will be prioritised in approvals. A broad coalition — the Clean Energy Council, ACF, WWF-Australia, the ETU and others, coordinated by the Carbon Zero Initiative — has gone further, insisting new facilities be powered by 100% additional renewable energy rather than certificates from existing projects, warning that social backlash is otherwise inevitable. Water is the next front: industry water demand is projected to more than triple from its current 5.5 GL.
The message to boards and investors is consistent: a data centre financed without a credible, additional, low-carbon power strategy carries disclosure risk, approval risk and social-licence risk — none of which sit comfortably in a financial model built on property assumptions.
From cost to value: the energy solutions that hold
Handled well, energy strategy turns from a cost line into the project's principal source of value and differentiation.
Each of these is a financial structure before it is an engineering choice. Each changes the risk, the cost of capital and the disclosure position of the asset.
Why this is a capital decision
The investment lens has already shifted. Data centres are increasingly underwritten like energy-infrastructure projects, with financiers assessing exposure to power markets — PPA terms, firming, hedging — as closely as rack density or throughput. Power strategy has become project value. Moody's estimates Australian data centres will need around A$10 billion of additional grid investment to support roughly 3.5 GW of extra load by 2035, on top of some A$50 billion of construction.
This is precisely where finance-led advice earns its place. The questions that decide outcomes are financial: Is the power position secured on terms that survive lender and investment-committee diligence? Is the sustainability commitment additional and defensible under AASB S2, or reliant on certificates that will not withstand examination? Does the capital structure reflect power-market exposure? Is the offtake creditworthy enough to unlock financing for the generation behind it? Answered poorly, these become stranded-asset and disclosure risks. Answered well, they become the investment case.
How Lever Impact advises
Lever Impact advises data centre and AI-infrastructure developers, operators and investors at the point where energy strategy, sustainability and capital meet. As chartered accountants working at the intersection of corporate finance and sustainability, we treat power as the starting point of value: structuring and stress-testing PPA and firming strategies as financial instruments, integrating Scope 2 and Scope 3 exposure and the Government's Expectations into AASB S2-ready disclosure, and building the capital structure and investment case that hold under diligence and earn a social licence. We do not sell the power or the assets — we make the decision defensible.
If you are developing, operating or financing data centre capacity in Australia and need the power strategy to hold — commercially, financially and under sustainability examination — let's talk. Book a consultation at leverimpact.com.au/book, or explore our energy transition, sustainability reporting and capital advisory capabilities.
For the full analysis, download our research paper: "Powering the AI Build-Out — energy, sustainability and the capital decision for Australian data centres."
Capital advisory and sustainable finance services referenced in this article are provided by Lever Impact Capital Pty Ltd (ABN 70 693 349 647), a Corporate Authorised Representative (AR No. 001319049) of BMYG Capital Pty Ltd (AFSL 505332), to wholesale clients only. This article is general information only, not financial product advice. Full disclosure →