What will we get out of the AI boom? The data suggests lots of noisy, energy-hungry datacentres and not much else
News Source
β’Wed, 26 Aug 2026 15:00:16 GMT
π° What Happened
In the Grogonomics column, Greg Jericho casts a sceptical eye over Australia's much-hyped AI investment boom, warning it is likely to produce plenty of noisy, energy-hungry datacentres but little real improvement in living standards. He points to official figures showing non-residential construction, which includes datacentres, has risen but is far from booming levels while wages have not taken off.
Although investment booms usually trigger economic growth, Jericho says this one shows few signs of trickling down to households. He revisits his analysis of March GDP, where he flagged that the AI tilt fails to translate into better living standards, and questions whether building datacentres on every vacant block really serves medical and scientific research.
π The Backstory
Australia has embraced a wave of corporate announcements about new hyperscale datacentre precincts, driven by cloud providers and global demand for computing power to fuel AI models. State and federal governments have courted large operators as engines of investment, framing the projects as essential to keeping the country at the forefront of AI.
Economists, though, are split on what datacentre construction truly delivers. Relative to the capital they absorb, datacentres employ few people per square metre and consume huge amounts of energy and water, so their multiplier effect on local jobs and wages can be weak. Jericho, long a sceptic of the AI productivity boom, argues the macro evidence confirms that headline GDP gains are not translating into prosperity for ordinary Australians.
π― Why It Matters
Communities and governments are pouring enormous attention into AI data centers. If such investment yields little in jobs, wages or living standards, it risks crowding out productive spending and repeating a boom-without-of-results pattern.
In the Grogonomics column, Greg Jericho casts a sceptical eye over Australia's much-hyped AI investment boom, warning it is likely to produce plenty of noisy, energy-hungry datacentres but little real improvement in living standards. He points to official figures showing non-residential construction, which includes datacentres, has risen but is far from booming levels while wages have not taken off.
Although investment booms usually trigger economic growth, Jericho says this one shows few signs of trickling down to households. He revisits his analysis of March GDP, where he flagged that the AI tilt fails to translate into better living standards, and questions whether building datacentres on every vacant block really serves medical and scientific research.
Australia has embraced a wave of corporate announcements about new hyperscale datacentre precincts, driven by cloud providers and global demand for computing power to fuel AI models. State and federal governments have courted large operators as engines of investment, framing the projects as essential to keeping the country at the forefront of AI.
Economists, though, are split on what datacentre construction truly delivers. Relative to the capital they absorb, datacentres employ few people per square metre and consume huge amounts of energy and water, so their multiplier effect on local jobs and wages can be weak. Jericho, long a sceptic of the AI productivity boom, argues the macro evidence confirms that headline GDP gains are not translating into prosperity for ordinary Australians.
Communities and governments are pouring enormous attention into AI data centers. If such investment yields little in jobs, wages or living standards, it risks crowding out productive spending and repeating a boom-without-of-results pattern.