Britain Will Fast-Track Your Data Center's Grid Connection. First, It Wants You to Earn It.

On October 28th, Chancellor John Healey stood up for his first Autumn Budget since taking over the job from Rachel Reeves in July, and buried inside it was a pitch aimed less at voters than at one industry: data centers. Britain says it needs at least 6 gigawatts of AI-capable data center capacity by 2030, roughly triple what exists today, and the thing standing between that target and reality isn't land or planning. It's the grid. Connection queues in the country's busiest markets routinely run past four years, and developers have wanted a workaround for a while. Healey gave them one, published the same day as a policy paper called the Data Centre Covenant, with a catch: you have to earn it.

The trade is simple to describe and harder to actually meet. A data center the government designates "strategic" gets two things it can't easily get on its own: priority placement in the grid connection queue, and permission to build its own on-site power generation while it waits for that connection to arrive. In exchange, it has to pass five tests covering local investment, jobs and UK suppliers, water and environmental standards, a fair share of the energy costs it creates, and security, specifically making compute available to the NHS and other public services. Every data center operating in Britain is "encouraged" to meet these tests. Only the ones asking for priority access actually have to.

The centerpiece is something the covenant calls a Local Benefit Plan. Strategic sites have to develop one with the communities actually living next to them, sized to match how big the project is, and stacked on top of the planning obligations, Section 106 agreements and the Community Infrastructure Levy, that already exist under English planning law. In practice, that's the government telling data center developers to formalize something American developers already do voluntarily when they need a township's vote: negotiate real money for things the town needs, write it down, and make it part of the deal. The difference is that Britain is making it a precondition for the one thing every developer actually wants, which is power.

What the covenant doesn't do yet is define its own terms. There's no megawatt threshold for what counts as "strategic." There's no stated minimum for how big a Local Benefit Plan has to be. The government says both are coming later. Until they arrive, the covenant is a framework without the numbers that would make it enforceable against any specific project, which is either a sensible way to avoid locking in a bad number on day one, or a sign the hard parts haven't been worked out, depending on how much benefit of the doubt you want to extend a government six months into a new Chancellor's tenure.

The planning system picks it up too

The covenant isn't standing alone. The government also published a draft National Policy Statement for Data Centres, open for consultation until January 8th, 2027, that would write the same five tests into the planning criteria used for the largest projects under the UK's national infrastructure regime. If that draft survives consultation, a data center big enough to count as nationally significant infrastructure won't face the covenant only as a condition of grid priority. It will face the same tests as a condition of planning approval, full stop. That's a meaningfully bigger lever than anything said at the dispatch box.

None of this is happening in isolation. The same Budget put more than £700 million toward AI investment generally, continuing a pattern of UK governments treating compute capacity as a sovereignty question as much as an economic one. Healey tied the covenant explicitly to that framing, linking priority grid access to keeping sensitive NHS and defense-related AI workloads running on infrastructure Britain actually controls, rather than compute Britain merely rents.

Industry likes it. The bill is still the problem.

techUK, the trade body most hyperscalers and colocation operators in Britain listen to, called the covenant an important recognition that domestic data center capacity matters for growth and national resilience, and welcomed a faster path through a grid queue that's held up projects for years. Its one real complaint wasn't about the five tests. It was about what the Budget didn't do: offer any relief on the electricity bills operators already pay, which techUK flagged as one of the weakest parts of the whole package. You can hand an industry a shorter line to the front of the queue and still not have answered the question it actually cares about most, which is the price per megawatt-hour once it gets there.

Developers have raised a more basic complaint: priority in a queue is only as useful as the substations and transmission capacity sitting behind it. Jumping to the front of a line that still has to wait on the National Grid to build physical infrastructure doesn't actually deliver power any faster, it just changes whose name is first on the list. The covenant speeds up who gets served first once new capacity exists. It doesn't, on its own, build a single new substation.

Voluntary, until someone tests it

The word doing the most work in the covenant's own text is "encouraged." Outside the strategic-site fast track, nothing in this document is mandatory, and even the strategic track only binds a developer who wants priority access badly enough to agree to the tests in the first place. Pinsent Masons, the law firm that tracks UK infrastructure policy for a living, flagged the obvious follow-up question in its own client note: what actually happens if the government doesn't deliver its side, the faster connection, the quicker planning decision, on schedule. The covenant doesn't say.

Britain has already seen what happens when a data center's community and environmental promises aren't legally binding. Earlier this year, the campaign groups Foxglove and Global Action Plan forced the government to concede, in a separate case over a hyperscale site in Buckinghamshire, that it had been wrong to approve the project without binding climate commitments attached. The developer ultimately accepted enforceable conditions, but only after campaigners sued, not because the system asked for them upfront. That's the precedent sitting underneath every "encouraged" in the new covenant: in Britain, much like in the US, a voluntary community promise and an enforceable one have historically needed a courtroom to tell them apart.

I think that's the real story here, more than the megawatt target or the £700 million. Every fight GridPermit tracks in the US over data center siting, the moratorium votes in Memphis and Oakland, the community benefits packages townships like Salem Township and Swatara Township negotiated project by project, the public-vote requirement Alabama's PSC just added to its own contract review, is the same argument Britain just tried to settle at the national level in one policy paper: should a developer get fast access to power and land without first proving, in writing, what the community actually gets in return. The US keeps litigating that fight one zoning board, one utility docket, at a time. Britain just wrote a national floor under it, tied directly to the thing every developer actually wants.

Whether that floor holds depends entirely on words the covenant hasn't filled in yet: what counts as "strategic," how big a Local Benefit Plan actually has to be, and what a community can do if a developer signs the covenant and then doesn't deliver on it. None of that is defined. Until it is, Healey has announced a direction, not a rulebook, and the first developer who tests whether "encouraged" means anything without a court order behind it will tell Britain, and everyone watching this model from outside it, whether the Data Centre Covenant is a real deal or just a very detailed press release.