New Cushman & Wakefield modelling suggests the answer will depend on how quickly adoption accelerates, and how its effects on jobs and productivity flow through the economy.
The firm’s AI Impact: National Insights Australia report models four scenarios and the potential impacts across major asset classes and capital cities.
We spoke to report author Sean Ellison, Cushman & Wakefield Director, APAC Research, about the findings.
Four Scenarios Test the Outlook
The report links changes in productivity, employment, GDP, inflation and interest rates to Cushman & Wakefield’s property forecasts to model four scenarios.
The baseline assumes gradual AI adoption, with modest productivity gains and steady economic growth.
A productivity-led expansion assumes faster AI uptake, stronger productivity and job creation.
In the AI bust scenario, AI companies’ revenues can’t cover the cost of the infrastructure build-out, producing a short, sharp recession in 2027.
The fourth scenario, displacement, assumes AI advances rapidly but increasingly substitutes for labour, pushing unemployment higher and weakening demand.
Mr Ellison said the scenarios were designed to show how different patterns of AI adoption could play out.
“We know that AI continues to be adopted at a rapid pace, but no one really knows what this will look like over the next one, three or five years. It really depends on how the technology evolves and how people use it,” he said.
“Imagining different scenarios sets goalposts for our clients and the industry, so they can see the different paths this technology might take and how each affects commercial real estate.”
Offices Show the Biggest Divide
Office markets are the most sensitive to the four scenarios, reflecting their exposure to knowledge-sector employment and demand for workspace.
The modelling also suggests AI’s effects will take time to emerge, becoming more visible after 2028 as businesses move from individual use towards broader enterprise adoption.
Under the baseline scenario, higher labour productivity is expected to support economic and employment growth from 2028 onwards.
Under the productivity-led expansion scenario, national prime office net absorption reaches 2.9 million square metres between 2025 and 2035.
Under the displacement scenario, it falls into negative territory, with net absorption of minus 419,000 square metres over the same period.
“AI affects knowledge workers in white-collar jobs, and those jobs are key to office demand,” Mr Ellison said.
“We went industry by industry, breaking occupations down into the tasks they perform and ranking those tasks according to how vulnerable they were, based on the available literature. That gives us a much clearer picture of where the potential impacts could emerge, and what they may look like.”
Industrial Maintains its Edge
Logistics and industrial property is more resilient across the scenarios, with demand for space continuing even as automation limits equivalent employment growth.
Under the baseline, national prime net absorption is forecast to reach 30.4 million square metres between 2025 and 2035.
Even under the displacement scenario, it remains positive at 27.8 million square metres.
Mr Ellison said industrial property had two factors working in its favour: greater flexibility in supply and the potential for growth in online spending to support demand.
“It’s quicker to put up a shed than an office, so the market can respond to changes in demand faster,” he said.
“In the displacement scenario, agentic AI means a greater share of purchases shifts online. Retail becomes more experiential, while online retail boosts logistics and industrial demand substantially. That shift from purchasing in-store to purchasing online helps offset some of the other effects.”
Cities Take Different Routes
The impact of AI also varies between cities, depending on their industry mix and exposure to sectors where AI adoption is likely to be greater.
Sydney, with its concentration of financial and professional services, is particularly sensitive to the different scenarios.
Melbourne is more resilient across the scenarios, supported by a broader economic base that includes professional services, education and advanced manufacturing.
“The makeup of Sydney CBD’s labour force makes it more sensitive to both the upside and downside risks presented in the scenarios than Melbourne, given the latter’s broad economic base,” Mr Ellison said.
In Perth, the strength of the resource sector could provide some protection from weaker conditions.
“Perth has a very large mining base, so even in the downside scenario, there is still demand for materials produced by the mining sector to support the regional AI infrastructure build-out,” Mr Ellison said.
“Australia is such a unique place. We have five or six big cities that are all very distinct, so modelling Perth isn’t the same as modelling Sydney, and Canberra isn’t the same as Melbourne.”
Bifurcation in Focus
For investors and developers, a key takeaway is the potential for an increasingly divided market.
“The main thing we’ve been talking about since COVID is increasing bifurcation, particularly in office markets,” Mr Ellison said.
The modelling also shows that the outcome will depend on how businesses use AI’s productivity gains - whether to protect margins or generate new demand.
History offers some cause for optimism, Mr Ellison said, with major advances in technology and productivity creating new demand, jobs and opportunities.
“It’s easy to see AI doing large parts of our jobs in the near future,” he said.
“What’s harder to imagine is what new tasks we’ll be doing, and what new jobs and opportunities AI will create.”
Cushman & Wakefield has created an AI Impact Barometer, which tracks the pace and spread of adoption and its potential impact on commercial property. View the tracker here.
