What the intake is actually for
Start with the composition, because the total is the wrong number to argue from.
Net overseas migration peaked at 530,700 in 2023, a post-border-reopening catch-up year that followed 438,000 in 2022. It has since fallen: 330,500 in 2024, 300,800 in 2025, a 43% drop from the peak.[^4] Anyone still writing about record-high migration is writing about a year that has passed.
The composition argument is different, and it holds across 2019, 2022 and 2023 regardless of the total. In 2023, students accounted for 51.7% of net overseas migration. Higher education alone was 33.9%. Permanent skilled migration, the category that addresses a genuine and lasting skills shortage, was 5.8% of the net figure. All skilled categories combined reached 14.1%.[^2]
So the intake is not primarily a skilled-worker intake. It is primarily a student intake, with a smaller skilled component, and a permanent skilled component that is smaller still.
That matters because of what automation does to the occupational distribution of the roles those students and temporary workers fill.
The mechanism: which work, done by what
Warehouse picking, food preparation, document processing, entry-level accounting, basic coding tasks, customer service routing, data entry. These are not abstract categories. They are the roles that appear most frequently in the temporary and student visa streams, and they are the roles where robotic process automation, large language model deployment and physical robotics investment is currently concentrated.
The argument is not that automation has already eliminated these roles. It has not, not at scale. The argument is that the capital commitment now underway will absorb most of them within the decade that the current intake will still be resident. The intake is permanent. The shortage it is sized for is temporary. That asymmetry is the structural problem, and it is the one neither Labor nor the Coalition has been willing to name.
The Productivity Commission named a related one in 2016: migration "does not offer a long-term panacea" for demographic pressure because "immigrants age too."[^3] The dependency ratio argument for high intake, the one that says Australia needs workers to support an ageing population, collapses on a long enough timeline. The automation argument collapses it faster, because it removes the labour shortage that was the premise.
Labor's record, in figures
The settings that produced the 2023 composition were Labor's. That is a record, not a characterisation.
The planning levels, the visa category weightings, the student visa conditions that allow full work rights during study and post-study work rights for up to six years: these are policy levers, and the government that held them set them where they were.[^5] The result was an intake in which permanent skilled migration, the category with the clearest economic justification, was outweighed roughly nine to one by students and other temporary entrants in net migration terms.
The per-capita GDP story sits alongside that. Headline GDP grew through a period in which GDP per capita fell. Adding people to the denominator is one way to grow an economy. It is not the same as raising the productivity of the people already here, and it does not produce the wage growth or the housing supply that makes the economy work for someone trying to enter it.
None of this is an argument about the character of any minister. It is an argument about what the settings produced, measured against what they were said to be for.
The Coalition's position, and why it needs a partner
The Liberals and Nationals now support reducing the intake. Angus Taylor has argued for lower immigration on cost-of-living and housing grounds. The Nationals have made the energy cost and regional infrastructure case.
The difficulty is that the Coalition presided over high intakes of its own in office. The 2018-19 net migration figure was above 230,000. The 2019 migration program planning level was 160,000 permanent places. These are not accusations. They are the record, and they are why the Coalition's current position reads as a change of mind rather than a demonstrated conviction.
That is not a reason to dismiss the position. People and parties update. But it is a reason why the coalition thesis, the one that has One Nation, the Liberals and the Nationals governing together at the next election, has One Nation leading on this specific question rather than following.
The Liberals bring the numbers and the ministerial depth. The Nationals bring the regional and resources mandate. One Nation brings the one thing neither has demonstrated in government: willingness to cut the net migration figure and hold it there against the lobbying pressure that always pushes the other way. Hanson has argued for lower intake consistently for longer than the current Labor government has existed. Joyce crossed to One Nation on 23 August. The leadership proposition is that the conviction is structural, not electoral.
The argument is not that the Liberals and Nationals are the enemy. They are the partners. The argument is that without One Nation holding the line, the line does not hold. History suggests that is right.
The blank space in One Nation's policies
One Nation publishes 29 policies on its website.[^6] None of them cover artificial intelligence, automation, data infrastructure or telecommunications.[^6]
That is not a gotcha. It is the most important policy brief in Australian politics right now, and it is sitting open.
Here is why it matters specifically for One Nation rather than generically for any party. One Nation's existing positions on foreign ownership, energy sovereignty and domestic industry point directly toward a set of conclusions about the data centre build-out that no other party has drawn yet.
Data centres are the fastest-growing new electricity load in Australia. They are capital-intensive, they require cheap and reliable power, and they are currently being built predominantly by offshore-owned entities. Australia holds the energy base, the land and the resource revenues to make domestically owned compute viable at scale. Norway turned its resource base into roughly US$1.8 trillion for 5.5 million people. Australia exports more resource value and has no equivalent fund. That is a policy failure, not a geological one.
A party that argues for domestic ownership of resources, cheap energy for Australian industry and reduced dependence on foreign capital is a party that has the building blocks of a compute sovereignty position. It has not assembled them. The Foreign Influence Transparency Scheme and the Register of Third-Party Lobbyists are public records. The question of who owns the infrastructure that Australian government AI runs on is a public interest question. These are the stories that sit inside One Nation's existing frame and have not been written as policy.
The section will argue what that policy should look like as the data comes in. The starting point is that a party about to be a governing force needs a technology policy before it governs, not after. The blank space is an invitation, and this section intends to fill it constructively.
The prize, not the warning
The migration argument is not primarily a warning. It is a precondition for something better.
If the intake is wound back, paced against measured displacement as automation absorbs the roles, the pressure on housing eases. The deposit-to-income ratio that has made homeownership a generational question rather than a personal one moves in the right direction. Graduate wages stop being suppressed by a labour supply that exceeds the demand automation has not yet removed. And the energy and resource base that currently funds other countries' sovereign wealth funds starts funding ours instead.
That is the argument for someone under 35 in Newcastle or Geelong who is working out whether their first job is the one that gets automated, whether their deposit will ever catch the market, and whether any party in Canberra has actually priced any of this in.
One Nation's record says it will cut the figure. The Productivity Commission says the dependency ratio argument does not survive the long run. The ABS data says the intake is concentrated in the most automation-exposed occupations. And 29 published policies say there is a technology brief waiting to be written.
That is the case. It rests on the record, not on character, and the record is in the sources below.