Could Australia’s AI Boom Make Our Housing Shortage Even Worse?

Housing has a new competitor.

Australia’s rapidly expanding data-centre industry is attracting billions of dollars, demanding substantial electricity and water capacity, and drawing upon many of the same engineers, consultants, contractors and financiers needed to deliver new communities.

The issue is not whether Australia should build data centres. It is whether governments are coordinating the digital, energy, infrastructure and housing pipelines or allowing them to compete blindly.

1. The capital is arriving quickly

The Reserve Bank estimates that Australian data-centre operators raised at least $35 billion during 2026 so far, compared with $24 billion throughout 2025 and an annual average of $4.6 billion between 2020 and 2024.

Debt supplied 85% of the 2026 funding. Data centres captured 17% of new syndicated lending, up from 6% during 2025.

The RBA says there is little evidence that this has materially affected financing conditions for other borrowers. That is an important qualification. Development finance is not presently being rationed because banks would rather fund server farms.

Why it matters: The speed and scale of capital formation show what housing is competing against. Data centres offer large loans, substantial collateral, long(ish) leases and creditworthy tenants (well for the time being). Residential developments offer planning risk, settlement risk, uncertain costs and buyers whose borrowing capacity is falling.

2. The bigger constraint may be infrastructure

AEMO expects data-centre electricity consumption to reach approximately 12 terawatt hours by 2030, that’s around 6% of grid-supplied electricity.

At the same time, Infrastructure Australia’s five-year major public infrastructure pipeline has reached a record $242 billion. Energy transmission, housing, transport and social infrastructure must all compete for design capacity, civil contractors, electrical specialists, equipment and materials.

This does not mean a data centre directly steals a carpenter from a townhouse project. The overlap is more likely to occur through engineers, power connections, substations, civil works, concrete, major contractors and government attention.

Why it matters: A residential site can possess approval and finance but still be delayed because electricity, water or specialist infrastructure cannot arrive when required.

3. Power availability becomes a property question

Power capacity has traditionally been treated as a technical due-diligence item. It is becoming a strategic land-use issue.

Developers should establish connection capacity, augmentation costs and delivery timeframes much earlier. Councils approving substantial residential growth should determine whether housing, industry and data centres are relying upon the same infrastructure upgrades.

The cheapest land may no longer provide lower home prices once connection costs and delays are included.

My two bobs worth

Data centres are not yet crowding housing out of finance, but they show where patient capital goes when projects offer scale, certainty and predictable income.

Housing offers almost the opposite: fragmented sites, changing rules, uncertain costs and retail settlement risk.

The immediate threat is not that a bank chooses a server farm over a housing project. It is that housing, energy, infrastructure and digital development are planned separately while drawing upon overlapping delivery capacity.

Residential feasibility studies should now test power availability, augmentation costs and infrastructure timing much earlier.

The next housing bottleneck may not be planning approval. It may be the connection queue.

About Michael Matusik Michael is director of independent property advisory Matusik Property Insights. He is independent, perceptive and to the point; has helped over 550 new residential developments come to fruition and writes his insightful Matusik Missive