Australia is building a lot of new homes. But is the datacentre boom holding us back? | Greg Jericho

What is a datacentre, and why are so many being built in Australia?

A datacentre is a specialised facility filled with computer servers, storage equipment, cooling systems and reliable power supplies. It processes and stores enormous amounts of digital information. Datacentres support online services, cloud computing and artificial-intelligence systems.

The article shows how rapidly this sector is expanding. Datacentres are counted within the Australian Bureau of Statistics category called commercial building not elsewhere classified. That category usually represents about 1% of non-residential building commencements by value. In the year to June, it represented 20%.

The article does not give a single cause for the boom. It highlights the massive resources going into datacentres and questions whether the benefits will reach anyone beyond AI companies selling “hopes and dreams”. The scale suggests strong investment in computing infrastructure, but it also increases pressure on builders, materials and capital needed elsewhere.

How much has Australian housing construction grown, and how large is the datacentre building boom by comparison?

Australian housing construction is growing strongly. The volume of residential building increased 8.7% in the year to June, the strongest result since 2016. Total building grew 10%, its best performance since the stimulus years following the global financial crisis.

Private housing commencements were also high. New private-sector house commencements rose 11.6% to 31,707 dwellings. That was the highest figure since 1994, excluding unusual Covid and pre-GST surges. Public and social housing remained much weaker, however, accounting for only 2.3% of new residential building.

The datacentre comparison is remarkable. Datacentres represented 20% of non-residential building commencements, compared with their usual share of about 1%. Their surge was worth just under three-quarters of the increase in non-residential commencements over the year. That makes datacentres a major rival for construction capacity.

Why could the rapid construction of datacentres make it harder to build enough homes?

Building homes requires scarce resources, including skilled workers, contractors, materials, equipment and investment. A datacentre boom does not automatically reduce housing construction, but it can intensify competition for those same resources. This matters when Australia is already trying to increase housing supply.

The article gives a clear example. Datacentres usually account for about 1% of non-residential building commencements by value. In the year to June, they accounted for 20%. Their construction was worth just under three-quarters of the increase in non-residential commencements. Builders and suppliers may therefore face stronger demand from this sector.

Housing construction is currently performing well, but public and social housing is not. It represented only 2.3% of new residential buildings in the year to June. Unless governments and industry expand capacity, datacentre investment could make it harder, slower or more expensive to deliver the homes Australia needs.

Why does Australia need more public and social housing, and how much of the housing stock does it currently make up?

Australia needs more public and social housing because it is facing a long-running housing crisis. The Senate committee reported a shared view that significantly more affordable and social housing must be built at scale. Government investment and clear delivery targets were identified as necessary.

The article describes a major gap between need and construction. Public and social housing has been barely part of the housing equation for more than 30 years. In the year to June, it accounted for only 2.3% of new residential buildings. Senator Barbara Pocock proposed a national target of at least 10% of Australia’s housing stock within a decade.

The article does not state the current percentage of the total housing stock. Its 2.3% figure refers to new residential building during the year, not all existing homes. That distinction matters: a small share of new construction can keep the overall public and social housing stock inadequate as demand grows.

How has the role of public and social housing in new construction changed since the period from 1955 to 1985?

Public and social housing once played a much larger role in Australia’s building program. Between 1955 and 1985, it represented an average of 13% of all new residential buildings. That made government-supported housing a meaningful part of overall supply.

The pattern changed after 1985. Over the following 30 years, excluding the global financial crisis stimulus period, public and social housing never accounted for more than 4% of new residential building. In the latest year to June, its share was only 2.3%.

This is a substantial reduction in the public sector’s contribution to new homes. It helps explain why the Senate committee called for more affordable and social housing built at scale through long-term government investment. Senator Pocock proposed increasing public and community housing to at least 10% of the country’s housing stock over the next decade.

What other uses of building workers and investment could compete with datacentres, such as homes or factories?

Construction workers, contractors, materials and finance can be directed toward many types of projects. Datacentres compete with housing, public and social housing, factories and other commercial buildings for this limited capacity. Choosing one project can delay or raise the cost of another.

The article offers a striking investment comparison. Normally, almost twice as much is spent building factories as buildings in the commercial not elsewhere classified category, which includes datacentres. In the past year, however, more than 17 times as much money was spent building datacentres as factories.

Housing is another competing use. Residential construction rose strongly, but public and social housing accounted for only 2.3% of new residential building. The issue is not that every datacentre replaces a home or factory. It is that a very large datacentre pipeline can absorb workers, materials and capital that governments and businesses might otherwise use to expand housing or productive industry.

Why can a shortage of builders, land, materials and capital limit how quickly a country can increase its housing supply?

Increasing housing supply requires more than planning permission or buyer demand. Projects need skilled builders and subcontractors, suitable land, materials such as timber and concrete, machinery and enough finance. If any of these inputs is scarce, construction slows, costs rise or fewer projects begin.

For example, a government might fund more social housing, but projects can still be delayed if builders are busy with datacentres. Limited land can restrict where homes are built. Shortages of materials can increase prices. Higher interest rates or insufficient capital can make projects too expensive to finance. These constraints interact rather than operating separately.

The article specifically warns that the available number of builders must be considered in housing policy. That warning is important during the datacentre boom, which accounted for nearly three-quarters of the increase in non-residential commencements. Faster housing growth therefore requires expanded construction capacity, not just stronger targets or incentives.

Key Facts:

📌 Datacentres house servers, storage equipment, cooling systems and power infrastructure.

📌 Their building category rose from roughly 1% to 20% of non-residential commencements.

📌 The article links the boom closely to AI companies and digital computing demand.

📌 Residential building volume rose 8.7% in the year to June.

📌 Private house commencements reached 31,707, up 11.6%.

📌 Datacentres supplied nearly three-quarters of non-residential building growth.

📌 Datacentres accounted for nearly three-quarters of non-residential building growth.