While actual numbers aren’t but accessible, CBRE believes synthetic intelligence (AI) companies have taken up greater than 1m sq ft of house in London between 2020 and 2025.
Network: Derwent London pre-let its 136,300 sq ft workplace growth in the West End to information and AI agency Databricks
More notable than the quantity, says Mike Gedye, the agency’s tech, media and telecoms sector lead, is the tempo at which AI corporations want to develop. “For speed, convenience and agility, they start in co-working because that’s the only place you can really get 15 to 20 seats quickly,” he explains. “But nearly even earlier than we’ve signed the lease or licence they’re saying we could also be 50 to 100 by the finish of subsequent 12 months.
“So suddenly they’re growing exponentially within flex and within a year they’re saying actually we probably need to be looking at more traditional spaces and then moving out from flex into traditional offices of 20,000 sq ft to 30,000 sq ft in 18 to 24 months. It’s incredibly fast.”
The use of versatile working house makes calculating the complete dimension of its footprint onerous as occupancy in the market is opaque.
In different instances, Gedye says, companies with solely 4 staff are searching for 20,000 sq ft as a result of they know they’re going to develop as they’ve secured enterprise capital backing.
AI-related companies have been making property headlines currently. This week, Derwent London pre-let its 136,300 sq ft Network workplace growth in the West End to information and AI agency Databricks. The deal adopted Great Portland Estates pre-letting 52,293 sq ft of workspace at The Delft close to London Bridge to AI software program agency Quantexa just a few weeks earlier.
Within a 12 months, AI corporations are shifting out from flex into conventional offices. It’s extremely quick
Mike Gedye, CBRE
Last week, Landsec reported that its new versatile workspace at MYO King’s Cross had seen a flurry of lettings led by AI and tech-led companies, with greater than 10 companies from throughout the technology sector taking house in lower than six months.
And tech giants are additionally in the combine, with OpenAI saying its intention to make London its largest analysis hub outdoors the US, though with out but detailing the dimension of its supposed workforce or workspace.
The progress follows a noticeable slowdown in the life sciences sector, which boomed simply after the Covid-19 pandemic however has since skilled rockier instances.
In November, British Land chief govt Simon Carter mentioned demand from AI companies had grown stronger than that of life sciences in the capital. Commentators say it’s now choosing up at a good sooner tempo than it was then.
Venture capital is driving the trade, says Tom Mellows, head of UK science, occupier illustration, leasing and growth at Savills, who likens the market now to that of life sciences earlier than its slowdown. “I think general economic headwinds have played a part, but also investors are quite short term,” he provides. “Lots of traders have realised that one thing like cell and gene remedy is kind of complicated, and it takes a very long time to show an concept right into a business product.
“AI you could argue is quicker, that its journey [to market] is quicker, in some instances, if it’s not related to something like a medical device or a new therapeutic treatment.”
AI companies embrace an enormous array of companies, Mellows provides, from cost platforms to these utilizing machine studying for enterprise effectivity. “It’s also quite often utilised as a tag because it helps raise money,” he says.
Some in scientific sectors similar to robotics, agritech and local weather tech require enhanced specs and both dry or moist lab house – which might be repurposed from earlier amenities supposed for all times sciences companies, in accordance with Mellows. Many others will simply be searching for customary offices.
Fast-growing companies
For landlords, the companies supply the pleasure of having a tenant from a fast-growing sector of their constructing. However, these occupiers can come with no confirmed observe document or a strong steadiness sheet, he provides.
Many scale-ups additionally need excessive ranges of lease flexibility because of their often-changing circumstances, which isn’t at all times engaging. “Very often, these fast-growing companies just don’t have a track record; they may not even have a covenant strength – you’ve got to be comfortable with that,” Mellows explains.
Gedye says landlords which have a mix of each versatile house and longer-term leases are nicely suited to money in on such companies.
The Delft: Great Portland Estates pre-let its workplace to AI software program agency Quantexa
British Land is a landlord that gives each. The agency’s head of campuses Mike Wiseman says the firm’s technique has at all times been round the broader science and technology space relatively than simply life sciences particularly.
The agency’s not too long ago redeveloped One Triton Square was initially positioned as a life sciences facility, however as of final week the 300,000 sq ft constructing was 77% let or below supply with a broader combine of occupiers. Biotech agency Gilead Sciences has taken 50,000 sq ft, however
British Land has hinted that AI corporations are additionally on board. Last summer season, Synthesia, one of the UK’s greatest AI corporations, moved into 20,000 sq ft headquarters at British Land’s close by 20 Triton Street.
Wiseman says there was “a decent chunk of deals” with “a lot more” house below supply. Although he says he can’t identify the corporations in query but, there’s a lot of “frontier tech demand coming through that building”.
And he feels London is completely positioned for additional progress in rising applied sciences because of its abilities base, academic establishments and the maturity of its tech firm choices. “There is a constant stream of people from the [US] West Coast particularly, saying to us that London is a super-attractive place to be, that the mature tech ecosystem is important because you have the talent here already,” he says.
London is the solely place in the world the place mature technology companies are based mostly so near cutting-edge analysis and growth, Wiseman believes.
Speaking to Property Week as he prepares to go to the US to fulfill with tech corporations, he says: “I’m going to see each these teams [researchers and established companies] after I go to the US, and I’ve bought a five-and-a-half-hour flight between Boston and San Francisco; that’s a good distance.
“I will also travel for longer between meetings within Boston than I do when I go up to see [semiconductor and software design company] Arm in Cambridge. I think we sometimes forget how well connected we are [in the UK].”
For this purpose, Wiseman thinks the ‘golden triangle’ of London, Oxford and Cambridge ought to be seen as one tech enterprise centre relatively than as separate areas.
According to Bidwells, there was 195,000 sq ft of house taken by AI companies in Cambridge at the finish of December 2025. Of these offering purposes, relatively than {hardware} or infrastructure, half of their work associated to the life sciences sector – one thing that highlights the complementary nature of the industries.
Bidwells analysis director Sue Foxley says: “There will be businesses better positioned than others, clearly, but the developments in the process of drug discovery and development, combined with the deep ecosystem in Cambridge, support the view that these activities will grow together rather than compete for space.”
But what about different areas of the UK? CBRE’s Gedye believes different cities with specialist industries and abilities may additionally turn out to be international hubs. He provides the examples of Bristol’s cybersecurity, Manchester’s drugs and science and Birmingham’s automotive and meals science companies.
“They’re going to have to become famous for something rather than trying to compete with London on a smaller scale,” he says.
Given the dizzying velocity at which AI-related companies have grown, discussions of the trade can’t escape the matter of whether or not there’s a bubble set to burst, particularly in gentle of the slowdown in the life sciences subject.
Mellows says: “It’s a bit like the dotcom bubble – you’re probably going to see ups and downs in investment because so much money’s got into the sector, but ultimately it’s here to stay: it’s going to be creating companies, creating demand in the long term and changing the way things are done.”