London Data Centre Outlook September 2026: Planning Is Strong, Power Is the Bottleneck

Quick answer

London’s data centre market in September 2026 looks busy on paper and constrained in practice. Planning activity is running roughly 20% above its 2021 to 2024 average, with a momentum reading of 119.5 against a baseline of 100. That is well down from the early 2026 peak of about 259, because the large filing wave of 2025 is now rolling out of the trailing twelve-month window.

The more important number is not the approval count. It is the gap between consent and electricity. City Hall data puts current peak data centre demand in London at around 760 MW, while the connection queue sits above 8 GW. That is more than ten times what the market actually draws today.

So the honest summary of the London data centre outlook for September 2026 is this: demand is real, planning is healthy, and delivery is the open question.

Where the London market actually stands right now

For most of 2023 and 2024, London planning activity ran below its historical norm. Then 2025 changed the picture. A cluster of qualifying schemes entered the system between March and December, pushing the momentum reading from under 60 to a peak in the high 250s by early 2026.

That peak has now faded. By June 2026, the reading had settled at 119.5. Nothing was cancelled to cause that fall. The older filings simply aged out of the rolling window faster than new ones arrived.

This matters because a falling index number gets misread as a cooling market. It is not the same thing. A trailing measure of new filings and first decisions tells you about the rate of fresh activity, not about the size of the pipeline already sitting inside the system. Both readings can be true at once: fewer new entries this year, and a larger stock of approved schemes than London has ever had.

What the planning momentum index measures, and what it does not

Anyone quoting a London planning figure should know exactly what it counts. The methodology behind the London Data Centre Outlook September 2026 index is deliberately narrow, which is what makes it usable.

The index adds one point for a project’s first qualifying planning filing and one for its first approval decision, then subtracts one for a first refusal or withdrawal. It runs on a trailing twelve-month window and publishes with a two-month lag.

Just as important is what sits outside it. Repeat documents for the same development do not create a second project. Minor equipment works and small refurbishments are excluded. Grid connection status and energisation dates are tracked separately. And nothing in the score implies that an approved scheme will ever be built.

Two design choices are worth calling out.

First, the two-month publication lag. Recent planning records arrive late and incomplete, so publishing the current month would manufacture a fake decline every single time. Holding data back until it matures is unglamorous and correct.

Second, the baseline is fixed rather than rolling. A score of 100 equals average net monthly milestones across 2021 to 2024, and the index is not capped. A reading of 240 means what it says. That is why the 2025 to 2026 wave is visible instead of being compressed into a ceiling.

The pipeline: 50 projects, 26 approvals, zero verified operational

Aggregate stage data from the August 2026 snapshot gives a cleaner picture than any single headline number. Fifty qualifying projects are being monitored, reconciled from 566 matching public planning records into one row per development.

The stage split breaks down as follows:

• One project at pre-consent
• Six still in planning
• Twenty-six approved
• One with verified construction evidence
• Ten recorded as build complete in the planning record
• None with independently verified operational status
• Six refused or withdrawn

One line in that list should stop anyone modelling future London capacity: 26 approved, one verified in construction, and none with independent public evidence of being powered and in service.

That is not a claim that London has no operating data centres. It clearly does, with City Hall counting 99 live sites. It is a statement about this specific planning-derived population, and it exposes how much distance sits between a committee vote and a functioning facility. Finance, procurement, construction, connection and commissioning all sit in that gap, and each one has its own failure mode.

Why power is the real constraint on the London data centre outlook

Planning consent is now the easier hurdle. Electricity is the hard one.

The scale of the mismatch is easy to state. London has around 99 operating sites drawing roughly 760 MW at peak. The reported connection queue exceeds 8 GW. Against that, the capacity where site, permission and power have all been reported as secured stands at about 1,289 MW.

A queue more than ten times current demand is not a forecast of ten-fold growth. It is a symptom of speculative queue positions, duplicate applications and optionality being held by developers who have not committed capital. Everyone in the sector knows it. The regulator now says so publicly.

Three developments in August 2026 tightened this picture considerably.

National Grid launched its North West London Upgrade on 13 August, covering a new substation at Letchmore Heath, around 60 km of new cable and more than 200 km of upgraded overhead line. The company links the programme to enabling new connections, including five data centres. Infrastructure of that scale does not get committed to speculative demand.

Ofgem proposed stricter tests for data centre grid connections, aimed at clearing capacity held by projects unlikely to proceed. If adopted, the surviving queue becomes smaller and considerably more credible.

JLL research showed hyperscale campuses moving away from established hubs. European hyperscale sites due to open within two years average roughly 175 km from major urban centres, against 46 km for those delivered between 2022 and 2025.

Read together, these point the same direction. Demand tied to London is not weakening. AWS added a fourth availability zone to its Europe (London) region in August, which is about as clear a demand signal as the market produces. But the physical answer to that demand is increasingly being built outside the capital’s traditional western clusters.

What this means if you are building, funding or approving

If you are a developer: consent is no longer your differentiator. A defensible connection route, and evidence of it, is. Sites with secured power will command a premium that has nothing to do with land value.

If you are an investor or lender: treat approval-stage and construction-verified assets as different risk categories, not different points on the same line. In this dataset, the ratio between them is 26 to 1.

If you are a planner or councillor: water use, carbon and community benefit are moving from soft considerations to material conditions. The Hounslow and Ealing decisions this summer both came with substantive scrutiny attached.

If you are a researcher or journalist: cite what a figure measures. Planning momentum is an activity index. It is not an approval probability, a capacity number, or a delivery score.

What to watch between now and early 2027

• Which schemes survive Ofgem’s proposed readiness and strategic-importance tests, and how quickly capacity is released.
• Whether the North West London Upgrade produces credible energisation dates for named sites rather than general reassurance.
• Whether emerging proposals outside the western clusters, such as the Charlton scheme reportedly offering around 180 MW across 100,000 square metres, mark the start of a genuine geographic shift.
• Whether any monitored project moves into verified operational status, which would be the first real test of the approval-to-delivery pathway.

Tracking these signals as they happen is now a live intelligence exercise rather than an annual reporting one. Platforms such as Authority.inc exist precisely because infrastructure markets move faster than quarterly research cycles, and because the difference between an announcement and a delivered asset is where most commercial risk sits.

Conclusion

The London data centre outlook for September 2026 is not a story of boom or bust. It is a story of two clocks running at different speeds. The planning clock is running roughly 20% ahead of its historical norm. The delivery clock is throttled by transmission capacity, connection queue reform and construction economics.

The approvals of 2025 and 2026 will not become computing capacity because a committee said yes. They will become capacity when power arrives. Anyone forecasting London supply should be counting megawatts energised, not permissions granted.

Frequently asked questions

What is the London data centre outlook for September 2026? 

Planning activity is around 20% above its 2021 to 2024 average, at a momentum reading of 119.5. The pipeline is large, but power access and delivery readiness are the binding constraints rather than planning permission.

Why has the planning momentum index fallen from its early 2026 peak? 

Because it is a trailing twelve-month measure. The heavy filing wave of 2025 has aged out of the window faster than new filings have replaced it. That reflects the timing of new entries, not cancellations.

Does an approved data centre in London mean it will be built? 

No. In the August 2026 snapshot, 26 monitored projects sat at approved stage while only one had verified construction evidence and none had independently verified operational status.

How big is London’s data centre grid connection queue? 

More than 8 GW, against current peak demand of roughly 760 MW. That queue includes speculative positions, which is why Ofgem has proposed tougher qualifying tests.

Where is new London-linked data centre capacity likely to be built? 

Increasingly at the edge of London and in secondary markets where land and power are easier to secure, even where the demand itself remains tied to the capital.

Is London still a viable data centre market? 

Yes, but the nature of the market is changing. Demand is strong and cloud capacity is still expanding in the region. The competitive advantage has shifted from securing consent to securing electricity.

What is the difference between build complete and operational?

Build complete means the planning authority recorded the completion of building work. Operational means independent evidence shows the facility is powered and in service. Only the second confirms usable capacity.