The mechanism nobody is costing
Large-language-model inference, the kind running inside every major enterprise software suite sold into Australia right now, requires sustained, dense electricity load. A single hyperscale AI data centre draws between 100 and 500 megawatts continuously. Several are under construction or in planning across New South Wales and Victoria. The International Energy Agency projects that data centre electricity demand globally will roughly double by 2030, and the Australian Energy Market Operator has flagged the accelerating commercial and industrial load from digital infrastructure in its own forecasting.
This is not a future risk. The grid investment to support that load is being priced into network tariffs now, and network charges are a component of every residential electricity bill in the National Electricity Market.
At the same time, the ABS Selected Living Cost Indexes show that other government transfer recipient households saw their measured living costs rise 51.7 per cent between the March quarter 2012 and the June quarter 2026. For pensioner and beneficiary households the figure is 50.0 per cent, and for age pensioner households 48.5 per cent. These are not CPI figures. The ABS Selected Living Cost Indexes include mortgage interest charges, which the CPI excludes, so they move differently and the two cannot be quoted interchangeably. They are index changes for a household type, not what any individual household experienced. But the direction and the magnitude are not in dispute, and they predate the data centre build-out accelerating into the grid.
The electricity component of those indexes is about to get a structural push from a load that did not exist at scale when the series began.
Australia exports the energy that could make domestic power among the cheapest in the developed world. It does not. Households and small businesses pay prices that reflect a market designed around export parity and a grid now absorbing industrial-scale compute load. The revenue from that resource base has not been converted into a sovereign wealth fund of the kind Norway built from oil for 5.5 million people. Australia exports more resource value and has no equivalent. That is not a grievance to nurse. It is a promise that has not been kept, and the people who cannot afford a deposit are the ones carrying the cost of that failure.
What the intake did to the infrastructure bill
Net overseas migration peaked at 530,700 in 2023 and has since fallen: 330,500 in 2024, 300,800 in 2025. The fall is real and it matters. It does not undo the infrastructure liability that the peak years created.
Population growth requires infrastructure delivery to match it. When delivery lags, the cost per head of existing infrastructure rises because the same assets serve more people, and the capital required to catch up is borrowed against future revenue. That is the mechanism behind a per-capita recession where headline GDP rose while output per person fell. The ABS overseas migration series documents the intake; the per-capita GDP divergence is the consequence that follows from it when supply cannot keep pace.
The Home Affairs migration program planning levels show the Family stream running alongside the Skill and Humanitarian streams. A permanent residency grant is not a one-off figure. It creates a future entitlement to sponsor family members, and the Family stream carries its own planning level. The standing population effect of a year's intake exceeds the headline net migration number, and a government planning only to the headline is planning to the wrong figure. The exact multiple is not something this section will invent. The direction is not in dispute, and it is documented in the program structure itself.
The Productivity Commission's answer to the dependency argument
The standard reply to any proposal to reduce the intake is that Australia needs migrants to support an ageing population. The Productivity Commission examined this directly in its inquiry into migrant intake into Australia and found that migration "does not offer a long-term panacea" because "immigrants age too." A policy built on continuous intake to manage the dependency ratio requires continuous intake forever, at an accelerating rate, to stay ahead of the ageing of the previous cohort. The infrastructure liability compounds with it.
That finding is from the Commonwealth's own productivity body. It is not a fringe position.
Why One Nation leads, and why the Coalition needs it to
The Liberal Party's cost-of-living plan acknowledges the pressure. Angus Taylor has spoken to reducing immigration. The Nationals have a genuine record on energy costs and regional infrastructure. These are real positions and this section argues for a coalition that includes all three parties.
The problem is the demonstrated record in government. The Coalition presided over high net overseas migration intakes during its own time in office. Labor set the settings that produced the 2023 peak: that is a record, not a characterisation, and it is documented in the ABS series. But the Coalition's current position on reducing the intake is a change of mind rather than a demonstrated conviction. A change of mind is reversible the moment a business lobby makes the call.
One Nation's proposition is not a trim at the margin. It is taking the intake down hard, year on year, continuing down, paced against measured automation displacement, and held there rather than reversed when the pressure eases. That is a schedule, not a slogan, and a schedule is what separates a governing commitment from a campaign position. A schedule can be modelled, costed and defended on air. "We would reduce it" cannot.
The Liberals and the Nationals bring the ministerial experience and the numbers in the lower house. One Nation brings the one thing neither has demonstrated in government: willingness to actually cut the figure and hold it there. That is the coalition thesis. One Nation leads on this question not because the other two are the enemy but because they have not yet shown they can hold the line without One Nation in the room.
What the schedule buys
A year-on-year reduction in net overseas migration, held consistently, produces measurable outcomes over a parliamentary term. Housing completions begin to outpace household formation rather than chase it. Infrastructure spend per head rises instead of being diluted across a larger denominator. Wage growth in the occupations most exposed to both automation displacement and labour market competition from the intake, entry-level administration, logistics coordination, junior accounting, becomes possible rather than theoretical.
These are not guaranteed outcomes from a single policy lever. They are the direction of travel when the lever is actually pulled and held, rather than pulled once and then allowed to drift back up. Both major parties have already run that pattern. The 2023 peak followed years of Coalition intakes that were themselves described as temporary responses to temporary shortages.
The technology policy gap
One Nation publishes 29 policies on its website. None of them cover artificial intelligence, automation, data infrastructure or telecommunications. That is a documented gap, not a characterisation.
It matters here because the data centre build-out is accelerating into the national grid right now, the electricity cost consequence is already in the pricing signals, and a party about to be a governing force will have to form a technology policy whether it intends to or not. The question is whether it forms one from its existing positions on foreign ownership, energy sovereignty and domestic resource use, which are strong foundations, or whether it inherits one from the vendors and lobbyists who will fill the vacuum.
The existing One Nation positions point toward Australian-owned compute running on cheap domestic power, with foreign investment in critical digital infrastructure subject to the same scrutiny as foreign investment in any other strategic asset. That is a coherent technology policy in outline. It has not been written yet. This section will argue what it should contain as the positions develop, because a governing party that has not priced the data centre load into its energy policy is not ready to govern on energy.