NextFin News - Datacentres already account for about 2% of the national electricity market in Australia, and Labor is preparing to set terms for datacentre and artificial intelligence growth as it tries to avoid repeating what it sees as the mistakes of the country’s resources boom, according to The Guardian on June 11.
That figure has shifted the issue from a niche policy concern to a mainstream economic one as AI-related infrastructure demand accelerates across the country. Datacentres are no longer being treated as simple real-estate projects for server racks. They are long-duration claims on electricity, land, water and transmission capacity.
Labor’s problem is how much of that load should be allowed to grow on commercial terms and how much should be tied to national-interest requirements. The resources boom is the reference point: export revenue rose, but so did concentration risk, infrastructure strain and the sense that the gains were unevenly shared.
Canberra’s language reflects that concern. Saying it does not want to repeat the resources cycle is not the same as opposing investment. It means the government wants to act earlier, before AI infrastructure is locked in by private decisions on grid access, site selection and local approvals. That could translate into closer scrutiny of where projects are built, how they source power, whether they can show efficiency gains and what commitments they make on jobs and community benefit.
The electricity demand alone explains why this has moved beyond a technology story. A sector using about 2% of the national electricity market is already large enough to affect generators, network operators and regulators, especially with demand still climbing. If AI adoption spreads faster than expected, the strain would reach beyond power supply to emissions targets, planning systems and the credibility of governments promising both digital growth and lower-carbon energy systems.
Labor’s position also shows how governments are starting to treat AI less as a pure productivity story and more as physical infrastructure with measurable costs. Datacentres use electricity continuously, unlike many industrial users, making them a sensitive addition to a grid already under transition. The question is not whether Australia wants AI investment. The question is whether it can capture the gains without subsidizing private expansion through underpriced infrastructure or lax approvals.
Australia has seen a similar cycle before. The mining expansion brought export income and investment, but it also deepened reliance on a narrow set of winners and drew criticism that policymakers moved too slowly on downstream diversification. Labor’s framing suggests it wants something different for AI: not a scramble to attract every possible project, but a framework that links growth to public goals.
There is also a risk in moving too cautiously. Datacentre operators usually need clear rules, fast approvals and confidence in power availability. If Australia adds conditions without improving the energy system underneath them, it could become a slower and more expensive place to build. That matters as global competition for AI infrastructure intensifies and capital moves toward jurisdictions offering scale, reliability and predictable regulation.
The case for intervention rests on where the costs and benefits fall. Datacentre growth creates concentrated costs and dispersed benefits. Local communities absorb the noise, land-use pressure and sometimes water demand, while technology vendors, cloud providers and large corporate tenants capture much of the upside. The government’s next test is whether it can turn that message into standards strong enough to matter without choking investment. The central fact remains the same: datacentres already consume about 2% of the national electricity market, and that number is rising.
Explore more exclusive insights at nextfin.ai.
