More Than 40 Data Centres, a Reported £30m a Year in Rates—and a £1m Ask: Who Controls Slough’s AI Benefits?

National infrastructure, local arithmetic

A town can host nationally important infrastructure and still ask where the benefit went. On 27 October 2026, the Financial Times reported that Slough has more than 40 data centres and that the borough council has invited operators to support projects ranging from £250,000 “visible quick wins” to £1 million in annual contributions for more transformative work. The same report cites an operator’s assessment that the sector contributes more than £30 million a year in business rates. Those figures should not be collapsed into one project account. The £30 million is a reported annual tax contribution, not a ring-fenced community-benefit fund, audited net local value or proof of what residents receive. The £1 million is a proposal described in current reporting, not money already committed, spent or delivered. Yet the contrast exposes a serious project-controls question: when national policy accelerates a portfolio of privately financed assets, who maintains the baseline for local value?

The trade-offs are unusually visible in Slough. The FT reports that council-owned land once intended for 1,000 social and affordable homes was sold to Equinix for £144 million in 2022, and that the developer later agreed a further £6.7 million package for training, transport and other measures as part of a 2025 planning consent. A capital receipt, a planning obligation, recurring business rates and affordable homes forgone are different categories, periods and counterfactuals; none should be added together as though they were comparable benefits. Context also matters. Slough’s 2026/27 budget report was based on £42.9 million of requested Exceptional Financial Support and £15.45 million of savings in that year. Those numbers do not prove that data centres caused the council’s financial position, or that their revenues could legally fill every gap. They do explain why residents may judge success by visible services, housing and opportunity rather than gross investment headlines.

Britain is accelerating what it cannot yet measure separately

Central government has made the national case forcefully. Data centres were designated critical national infrastructure in 2024; June 2026 cyber-resilience proposals would treat qualifying operators as essential services. The government’s AI Growth Zones delivery plan promises stronger planning support, a £4.5 million national team to help local planning authorities, land safeguarding and streamlined nationally significant infrastructure consenting. The direction is clear: Britain wants faster data-centre delivery to support AI, cloud services, public systems and economic growth. However, the Office for National Statistics said on 13 October 2026 that it cannot yet identify data-centre investment or output separately in the national accounts. That is not an argument to stop building. It is a warning against managing a fast-growing infrastructure class with a benefits vocabulary more precise than the measurement system behind it.

The June 2026 ministerial decision on Manor Farm in Slough illustrates the scale and the contest. Evidence reproduced in the inspector’s report described 236MW of past data-centre completions, including 206MW on the Slough Trading Estate over six years, and a potential 559MW estate pipeline over the following seven years. It also referred to 4.3 million square feet of possible data-centre development under the estate’s simplified planning framework. These are evidence and capacity propositions in a planning case, not guaranteed operating output, investment, jobs or public benefit. Project controls should therefore distinguish permission, power capacity, construction progress, operational compute, tax receipts and local outcomes. Calling them all “growth” creates a dashboard that can remain green while the public argument turns red.

Academic evidence: legitimacy and structure must move with the project

A new open-access Project Management Journal paper gives the dilemma a useful name. Lavagnon Ika’s You Can’t Always Have It Both Ways, first published on 23 September 2026, argues that accelerating a megaproject may strengthen pragmatic legitimacy—people see momentum and utility—while weakening moral legitimacy if fairness, participation or accountability appear compressed. Stronger scrutiny can reverse that relationship while also slowing decisions. Ika’s response is adaptive governance: leaders should prioritise, sequence, recombine and reinterpret responses as conditions change, rather than pretend that speed and accountability can both be maximised at every moment. This is a conceptual “thoughtlet”, not an empirical evaluation of Slough or data centres, so it cannot establish that one planning route produces better outcomes. Its value is diagnostic: the local question is not an inconvenient communications risk outside the programme; it is part of whether the infrastructure remains publicly legitimate.

Empirical evidence published in the International Journal of Project Management in July 2026 strengthens the organisational implication. Kaiwen Jiang and colleagues’ open-access study of four major airport programmes in China, representing almost US$20 billion of investment, drew on participant observation, archival records, 94 semi-structured interviews and 308 observed meetings. It found that dominant uncertainty shifted across the lifecycle—from ambiguity in planning, through volatility in delivery and complexity at integration, to criticality in operations—and that client organisations adapted by changing where decisions were centralised and how responsibilities were grouped. The researchers’ schedule-and-risk consultancy role gave unusual access but created potential bias, which they addressed through document triangulation, multiple analysts and member checking. Transfer to England must remain cautious: four airports in China’s centralised institutional setting do not prove what Slough Council or a private data-centre operator should do. They do support a practical principle: governance designed for planning permission should not remain frozen when the problem becomes cumulative infrastructure, operations and community value.

Give local value its own control account

For Slough, operators and national government, I would create a portfolio-level local-value control account alongside each facility’s cost and schedule controls. Start with four baselines: land and alternative use; power, water, heat, transport and environmental capacity; public finance; and community conditions. Then separate measures into inputs, outputs and outcomes. A signed contribution is an input. An apprenticeship place offered, a bus service funded or waste heat connected to homes is an output. Sustained employment, lower household cost, better access to services or reduced inequality is an outcome—and it needs a denominator, comparison period and credible account of other causes. Record who benefits, where, when and for how long. Publish forecast, commitment, payment, delivery and realised outcome as separate fields. This would prevent a £144 million receipt, £30 million of annual rates and a £1 million proposed contribution from being presented as interchangeable evidence of success.

Governance should also change as the portfolio changes. During planning, the council and developer need transparent alternatives, cumulative-impact evidence and a clear community proposition. During construction, operators and contractors should report local supply-chain access, safety, disruption, skills and delivery against obligations. Before operation, an integrated readiness review should cover grid dependencies, cyber resilience, water, cooling, emergency arrangements, jobs and monitoring data. Once operating, a standing benefits forum should bring together the council, operators, utilities, colleges and resident representatives; it should have decision rights, not merely receive presentations. National bodies should fund the specialist capacity demanded by nationally significant decisions, while operators provide comparable facility data without exposing security-sensitive information. A quarterly exception report should ask not only “is the facility on programme?” but “which promised local outcome is off trajectory, who can act, and by what date?” These are original recommendations, not commitments made by Slough Council or any operator.

Slough’s position can become an advantage. Few places have comparable experience of hosting the physical estate behind Britain’s digital economy, and that experience can inform better national practice. The positive choice is not between stopping AI infrastructure and giving it a blank cheque. It is to join delivery speed with visible accountability: define the local benefits before approval, preserve the counterfactual, adapt governance as uncertainty changes, and report what residents actually receive. More than 40 facilities constitute a portfolio, not a series of isolated planning files. Managed as a portfolio, they can support national resilience, skilled work and stronger local services while earning the public confidence on which future investment depends. The next board should ask a question every major infrastructure sponsor can use: if this project succeeds technically, what evidence will show that its host community succeeded with it?