AI Is Becoming a Location Filter for European Commercial Real Estate
Artificial intelligence is often discussed as a space reducer. The sharper signal is different. New research from Cushman & Wakefield suggests AI may support commercial real estate demand across Europe, not by creating a blanket need for more floor area, but by changing which buildings qualify as useful, investable and resilient.
That distinction matters for anyone reading property markets through data. The next decade is unlikely to be measured only by total occupied square metres. It will be measured by the spread between buildings with power capacity, connectivity, flexibility, sustainability credentials and transit access, and those without them.
Cushman & Wakefield’s EMEA research uses econometric modelling and scenario analysis to test how AI adoption could affect productivity, employment, occupier demand, data centre growth, regulation and real estate values. The important analytical point is that the model does not treat AI as a single outcome. It frames AI as a probability set, with different implications under gradual adoption, rapid adoption, underperformance or sharper labour displacement.
In the baseline case, AI improves productivity and supports economic expansion, which feeds into occupier activity. In the upside case, faster adoption creates stronger growth, new business formation and higher real estate values. The downside cases are equally important. If AI disappoints or displaces labour faster than markets can absorb, vacancy risk rises and rents come under pressure. For investors, this makes single-point forecasting less useful than scenario-weighted underwriting.
The office market is where the signal is most easily misread. AI can reduce demand linked to routine tasks, but that does not automatically mean weaker office demand overall. It can also create new firms, new roles and more demand for collaboration, client engagement and decision-making environments. The likely result is not a simple contraction. It is a quality migration.
AI will not raise the value of every square metre. It will make the difference between capable and obsolete space more measurable.
For property intelligence teams, this shifts the core dataset. Traditional indicators such as vacancy, rent and leasing volume remain essential, but they are no longer enough. Buildings need to be scored against energy availability, digital infrastructure, grid constraints, refurbishment potential, amenities, public transport access and carbon performance. AI demand is not just an occupier trend. It is an infrastructure test.
Logistics assets face a similar filter. AI-enabled supply chains reward facilities that can handle automation, faster throughput and more complex inventory decisions. Generic sheds in weaker locations may not capture the same uplift as modern, flexible assets near labour, transport and power. In data centres, the connection is more direct. AI workloads intensify demand for compute capacity, making grid access and power reliability central location variables rather than technical footnotes.
Retail shows the most selective pattern. AI is unlikely to create broad demand for additional retail floorspace. Instead, it may sharpen the divide between prime, experiential, service-led locations and generic mid-market space. Better customer analytics, inventory systems and personalised commerce can strengthen the best physical retail, while exposing weaker propositions faster.
The practical takeaway is clear. Owners and occupiers should track AI exposure at asset level, not only at market level. The key questions are which buildings can support higher digital and energy loads, which locations attract AI-driven firms and talent, and which assets become less competitive as occupier requirements rise. AI is not replacing the real estate cycle. It is adding a new layer of intelligence to it.
Source: Cushman & Wakefield
