Where are AI companies moving to in London?

King’s Cross is the primary destination, anchored by Google and OpenAI and supported by the wider Knowledge Quarter. Most other activity sits within the King’s Cross to City to Shoreditch and Old Street triangle, with Soho emerging for AI businesses with a creative or consumer-facing angle.

Should an AI startup take a traditional lease or a flexible workspace in London?

It is often the case that Flex is predominantly used up to 10,000 to 15,000 sq feet and where headcount is difficult to predict in the medium term. High-growth AI businesses are typically better served by serviced or managed offices because the optionality outweighs the per-foot saving on leasehold. Some Cushman & Wakefield AI clients remain in flex at 25,000 sq ft and above.

At what size does traditional leasehold become worth considering for an AI company?

Around 10,000 to 15,000 sq ft is the threshold at which leasehold becomes worth modelling seriously. The decision depends more on growth trajectory and forecast visibility than raw size. If you can see your headcount five years out with confidence, leasehold becomes interesting. If you can’t, flex usually wins.

What do AI and tech occupiers prioritise when choosing a London office?

The five recurring requirements we see are: flexible terms, on-site amenities that help attract talent, speed of delivery and activation, expansion options and rights of first refusal in the same building, and access to additional meeting rooms during busy periods such as fundraising or hiring sprints.

AI Companies Moving to new offices in London: where to search, what benefits Flex can offer

If you’re an AI company moving to or expanding in London, here’s the short answer:

Where: King’s Cross is the gravitational centre of London’s AI scene, anchored by Google and OpenAI. The wider triangle of King’s Cross, the City and Shoreditch/Old Street covers most of the activity, with Soho and Paddington emerging fast.

How: Flexible workspace, serviced or managed offices, fits a lot of the AI businesses we advise launching in London. A five to ten-year lease is hard to commit to when you might be 30 people today and 200 in eighteen months, and that flexibility on term length and space take is why flex appeals. Some AI clients choose it even at 25,000 sq ft and beyond.

Flex or a traditional lease? It depends on your situation: the type of arrangement, the size you need, the location. A conventional lease is the right answer in plenty of cases. We cover both and stay agnostic, so the honest short answer is to talk to our team and pressure-test the options against your actual requirements.

Why London, and why now

I run Flexible Office Advisory UK at Cushman & Wakefield. I spend my week in the room with founders, ops leads, and US-headquartered AI firms working out where in London they should land, what kind of building fits how they actually work, and how to structure the deal so they’re not boxed in eighteen months later.

We’re seeing a genuine wave of AI companies, frontier labs, applied AI scale-ups, US firms opening their first European HQ, all mobilising into London at the same time. The pull factors are familiar but the pace is new: the talent pool from UK universities and the existing tech ecosystem, transport links into Europe by train and plane, and the gravitational effect of the firms already here. Once a Google, Anthropic, Open AI or others plant a flag, every other AI business wants to be within walking distance.

This guide is what I’d tell you on a first call. No per-desk pricing, no listings, just the decisions to be made.

Where AI companies are basing themselves in London

King’s Cross: the Knowledge Quarter and London’s AI heart

King’s Cross is the centre of gravity. Google’s UK headquarters sits there. OpenAI is there. The Knowledge Quarter, which clusters King’s Cross with Bloomsbury and Euston, holds an extraordinary density of universities, research institutions and AI-adjacent businesses. If you’re hiring AI engineers, researchers or product leads, this is where the talent already is, and where it wants to come for coffee.

Transport is the second reason. King’s Cross and St Pancras together give you the East Coast Mainline north, the Eurostar straight into Paris, Brussels and Amsterdam, six Tube lines, and easy runs to Heathrow, Stansted and Gatwick. For US firms running a London HQ that needs to plug into European clients and investors, that matters more than founders sometimes realise on day one.

The current AI triangle: King’s Cross, the City, Shoreditch and Old Street

Historically, the London tech story was Old Street and Shoreditch. The “Silicon Roundabout” branding stuck for a reason. Those submarkets are still where a lot of earlier-stage tech and AI startups land first, and the buildings are well suited to it: characterful, well-served by flex operators, with the kind of amenity stock that attracts talent.

What’s shifted is the shape of the map. The active AI triangle now runs King’s Cross to the City to Shoreditch and Old Street, with Farringdon stitched into the middle by the Elizabeth Line. The City has reinvented itself for occupiers it would have ignored a decade ago, and there’s a lot of high-quality flex and managed product available there now.

Soho: the emerging story

Soho is becoming an interesting alternative for certain AI and tech sectors, particularly those with a creative, media or consumer-facing angle. The buildings are smaller, the amenity is genuinely walking-distance, and there’s a different kind of energy to it than the Knowledge Quarter. Worth shortlisting if your team would rather be among film, advertising and fintech than next door to a research lab.

Layout follows function, not company type

There is no such thing as “an AI office layout”. The layout you need depends on what your team actually does in the building. A research-heavy team with deep-work needs looks completely different from a sales-led team running enterprise demos all day. Don’t let an agent or operator sell you a generic “tech fit-out”. Tell them how your people work, and design from there.

Why flexible workspace is winning for AI scale-ups

Tightly constrained long term leases are often incompatible with high growth AI businesses. The cost of getting it wrong, signing for too much space, too little space, or the wrong building, is high enough that the security of a long lease becomes a liability rather than an asset.

Flex, whether that’s serviced offices, managed offices, or hybrid models, gives you the optionality to scale up, scale down, or move entirely without taking a financial hit you can’t recover from.

These are five things that we often hear in briefs and requirements from AI and growing tech clients.

1. Flexible terms that don’t lock them in beyond their funding visibility.
2. High end amenities on-site, end-of-trip facilities, café, events space, the things that help them attract talent.
3. Speed of delivery and activation. They need to be operational in weeks, not quarters.
4. Expansion options and rights of first refusal on neighbouring floors, so they can grow inside the same building rather than moving every twelve months.
5. Meeting room access for busy periods, particularly when they’re running fundraising rounds, partner days or hiring sprints.

The decision isn’t really about square footage. It’s about growth trajectory and visibility. If you can see, with reasonable confidence, what your business looks like in five years, leasehold becomes interesting. If you can’t, flex is almost always the right answer.

How a Cushman & Wakefield Locate London advisor helps

We work as advisors, not brokers. That distinction matters. We sit on your side of the table, run side-by-side comparisons of serviced, managed and leasehold options, and stress-test them against your funding plan, your hiring plan and your three-year operating model. Then we negotiate.

For most AI scale-ups, that process saves significantly more in lease terms, fit-out contributions and incentive packages than it costs, often it costs nothing at all on flex. And critically, it stops you signing the wrong deal in the wrong building, which is the most expensive mistake an AI founder can make in their first London move.

If you’re early in your search, the best thing to do is talk to us before you’ve shortlisted buildings. Once you’re walking properties, your options have already narrowed.