It seems the picture emerging from the UK’s leading innovation markets is far more complex. AI isn’t reducing demand; it’s redistributing it. It’s concentrating activity in places where specialised science, proprietary data, and deep technical talent intersect. And it’s raising expectations for what an office must be.
This was the clear message from today’s panel featuring Bidwells’ Research Director Sue Foxley and Max Bryan, Head of Science & Technology, along with Jane Hutchins from Cambridge Science Park, Pioneer Group’s Toby Reid, Gareth Fisher of ATOS and Stephen Muclair from 3RE Capital Ventures. Their collective view: AI is reshaping the office market, but not in the way many predicted.
AI as an Amplifier, Not a Substitute
The instinctive narrative, with fewer people, fewer desks, doesn’t hold up when you look at what’s actually happening on the ground. As Sue Foxley put it, “Demand is not softening – it is structurally reallocating.” AI is acting as an amplifier in sectors where human expertise remains essential. Productivity gains don’t eliminate work; they create new categories of work, new workflows, and new spatial needs.
The panel described a market splitting into three broad groups. At one end are the AI natives - companies like OpenAI and Anthropic - whose growth is explosive and whose space requirements are highly specialised. London dominates this category, with 72% of the UK’s AI-native companies and 90% of the employees. These firms are taking large, premium spaces in central locations, and they are doing so quickly.
Then there are the innovation-led businesses: Life sciences, drug discovery, robotics, genomics, where AI accelerates discovery but doesn’t remove the need for labs or engineering space. Cambridge, Oxford, Manchester and Liverpool all sit in this category. These ecosystems are “geographically sticky,” as the panel described them, because talent, research infrastructure and collaboration networks can’t be replicated elsewhere.
Finally, there are the professional services and admin-heavy sectors. This is where AI-driven substitution is most visible. Some roles will shrink. Some functions may consolidate into London. And cities without strong innovation identities will feel the pressure most acutely.
What AI-Era Occupiers Want from Buildings
Across all sectors, the panel agreed that occupiers are demanding a fundamentally different kind of building. Power and digital resilience are becoming as important as location. AI workloads require more electrical capacity, more cooling, more fibre, and more secure compute environments. Grid capacity is becoming a strategic issue, and in some cases, a competitive advantage.
Flexibility is another defining theme. AI changes workflows quickly, and buildings must be able to adapt. Floorplates need to support shifts between office and lab space. Ceilings need to be higher. Risers need to be oversized. The days of rigid, single-purpose offices are fading fast.
But perhaps the most striking shift is cultural. Despite the hype, AI doesn’t replace creativity—it amplifies it. And creativity thrives in human environments. Occupiers want places where people can bump into each other, scribble ideas on walls, grab coffee, attend events, and feel part of a community. As one panellist said with refreshing bluntness: “Teams doesn’t work… in the creative environment.”
This is why Cambridge Science Park, for example, is investing as much in its social infrastructure as its physical one. Talent attraction and retention now depend on the quality of the environment as much as the quality of the job.
The Geography of AI: Winners and the At-Risk
The geography of AI-driven demand is becoming clearer. London, Cambridge and Oxford are structurally advantaged. They have the talent, the capital, the research base and the global visibility. Specialist regional hubs such as Manchester, Liverpool and Newcastle, also have strong foundations, provided they double down on their defensible strengths.
The places most at risk are those reliant on process-heavy professional services and those with large volumes of secondary office stock that cannot be easily upgraded. The panel was frank: trying to be a generalist city in an AI world is a losing strategy. Specialisation is the only sustainable path.
A Market Splitting in Two
Developers face a new reality. The market is bifurcating. Prime, high-spec buildings are thriving. Secondary stock is struggling. One speaker described it as a “Champagne Tower”: the top glass keeps filling, but the lower ones are running dry.
Planning and power constraints are becoming central to investment decisions. Flexibility is no longer optional. And soft infrastructure, namely culture, community, events and networking, is now part of the value proposition.
So…Will AI Increase or Decrease Office Demand?
The panel’s answer is both simple and subtle. AI will increase total office demand in the UK, but in a highly uneven way. There will be more space for innovation clusters, less for generic office markets, and much higher specification requirements everywhere.
The question is no longer how much space we need. It’s where, what kind, and for whom. As Sue Foxley summarised, “It’s not that AI is going to take away the need for office space… the challenge is what happens where.”
