Opinion | Policy & Power
The Budget’s Data Centre Covenant shows the government has understood that power is the constraint, but a faster route through the grid queue is not the same as more electricity – and the detail that matters is still to come.
Power is finally on the policy agenda
For much of the past year, UK developers have argued that the binding constraint on AI infrastructure is not capital, land or demand. It is time to power. Reports that Microsoft has pulled out of Nscale’s Loughton project after grid officials reportedly said power would not arrive until the 2030s gave that argument a headline. Microsoft has not commented, and no reason was given.
The Chancellor’s Data Centre Covenant announced at the Autumn Budget yesterday is the first serious attempt to respond. The government estimates the UK will need 6GW of AI data centre capacity by 2030, and it has put grid access at the centre of the offer and whilst that deserves credit, it also deserves some scrutiny.
What exactly is on offer, and what is asked?
The mechanics are worth stating precisely; A new Ministerial Control Centre, chaired by the Minister for AI, will use state power to speed up time to power and planning for the most strategic data centres. For sites that qualify, the government will reserve future grid capacity, reallocate capacity released when other projects leave the connections queue, and facilitate transfers of pre-2030 connection contracts to strategically important projects.
In return, five tests cover electricity costs, local investment, environmental standards, jobs and apprenticeships, and security. All data centres are encouraged to meet them, but strategic ones must do so to qualify for the extra support. The environmental test includes a Power Usage Effectiveness of 1.3 or better for new facilities and closed-loop water technologies, and strategic developers must agree a Local Benefit Plan linked to the size of the development. The Chancellor framed this around sovereignty: “When the NHS uses AI to treat patients, we want that data here,” he said. Keeping critical workloads in Britain is a defensible aim, the question is whether the policy can deliver it on the timescale it sets.
An industry welcome, with conditions
The initial reaction from the infrastructure and energy sectors has been constructive. In a statement, Matthew Baynes, Vice President, Secure Power and Data Centres, UK&I at Schneider Electric, framed the Covenant as an opportunity:
“The Chancellor’s Data Centre Covenant represents a significant opportunity for a new era of responsible datacentre development, where investment in digital infrastructure delivers tangible benefits for local communities and the wider UK economy. As AI accelerates demand for computing capacity, the industry has an opportunity to become a catalyst for economic growth, creating skilled jobs, attracting investment and strengthening the UK’s position as a leading hub for secure, scalable and sovereign digital infrastructure. However, realising this ambition will require a fundamental rethink of how we power the next generation of data centres.”
That final sentence is the one that matters – Schneider Electric supplies power and cooling systems, so Baynes has a commercial interest in the answer, but his diagnosis is shared across the sector.
Giles Hanglin, CEO of Apatura Energy, went further and picked out the delivery mechanism in his own statement:
“At Apatura, we welcome the new proposals from the Chancellor, and believe that the new Ministerial Control Centre, with its proposed powers to reserve and reallocate grid connections, is the strongest measure set out in today’s Budget. Time to power remains the single biggest constraint on UK data centre investment. The speed of delivery, alongside the cost of energy, will be the deciding factors in the country’s economic growth.”
Not everyone is persuaded though – the CEO of Greenpixie dismissed the Covenant as a smokescreen to appease local resistance to AI infrastructure.
Reallocation is not generation
The strongest feature of the Covenant is also its most limited; Reserving and reallocating connections moves capacity from one project to another, it doesn’t add any. That is useful if the queue is clogged with speculative applications and the government is backing Ofgem’s commitment fee and queue management measures to deter speculative demand. Clearing dead wood has real value.
The sums matter. Prioritising one project means de-prioritising another, and the 6GW estimate still has to be met by physical connections – One think tank has pointed out that the government’s own preliminary estimates suggest the UK could reach as little as 3.3GW, with viable projects stuck in the grid queue. If that is right, then reallocation narrows the gap, but it does not close it.
Hanglin’s own warning points to the risk:
“Where the risk lies is in the detail still to come. The definition of a strategic data centre, the criteria the Ministerial Control Centre will apply and the timetable for its decisions are all left to further announcements, and if that takes months the fast lane will not be fast. Our ask is that government publishes those criteria quickly and judges projects on readiness and community contribution, so that the sites able to deliver this decade are the ones that get the power.”
It is the clearest statement of what the industry needs next – a control centre with powers to reallocate connections is only as good as the criteria it applies, and those criteria do not exist yet.
Bridging power is the pragmatic measure, but a narrow one
The most practical element may be temporary on-site generation, which developers have argued for months. Baynes, whose company supplies this kind of technology, was explicit about why it matters:
“The commitment to prioritising grid connections for critical data centres, alongside enabling on-site power generation where necessary, is particularly encouraging. With power availability increasingly influencing where and how new facilities are developed, behind-the-meter solutions, including microgrids, battery energy storage and alternative generation technologies, can play a pivotal role in accelerating deployment, improving energy resilience and reducing pressure on the grid.”
The caveat is in the small print: The framework will be tightly controlled, available only to exceptionally strategic projects, only where the other measures cannot secure capacity in time, and subject to conditions such as energy security. The same think tank argues that this permission should be opened up more widely, with clear objective criteria rather than ministerial discretion. That seems right – if only a handful of projects can build their own bridge, then the many viable ones in the queue stay where they are.
The delivery stack is getting heavier
There is also the cumulative burden.
A strategic developer will face planning requirements, Section 106 and the Community Infrastructure Levy, the Covenant’s tests, a Local Benefit Plan and annual reporting. Legal advisers have already asked whether the Covenant sits outside the planning system or is imposed through it, and warned against overlapping or conflicting requirements.
None of this is unreasonable in isolation. Communities should share in the benefits of hosting critical infrastructure, but every added obligation is a negotiation, and negotiations take time and whilst the Ministerial Control Centre can accelerate decisions – It cannot negotiate local agreements on a developer’s behalf.
What the government should do next
The Covenant is more than a headline – It sits alongside wider reforms: greater freedom for developers to self-build connection infrastructure, more flexible connections and a draft Data Centre National Policy Statement. The direction is right, the risk is sequencing.
Three things that would turn the framework into delivery:
- Publish the definition of “strategic” quickly, with a megawatt threshold and objective criteria. The government has said further details will follow in due course.
- Judge projects on readiness. Sites with planning, finance and a credible connection date should come first, not the ones with the best announcements.
- Widen the bridging-power route, with clear conditions, so more viable projects can build while they wait.
The test that matters
The Covenant is a welcome recognition that digital infrastructure is critical national infrastructure – Baynes captures what has to happen next:
“The opportunity now is for government, energy providers and the data centre industry to work together to turn this commitment into action. By embracing innovative energy technologies and taking a more integrated approach to infrastructure planning, we can unlock the next wave of AI investment while ensuring data centre growth is sustainable, resilient and delivers lasting value to the communities it serves.”
Policy can only allocate what the system can physically deliver – the true measure of success will not be how many projects are designated as strategic – it will be how many megawatts are energised by 2030.
