Lammermuir arithmetic: £2bn of data centre against 10,000 signatures, before an application exists

Ten thousand names arrived before there was an application to read. The Lammermuir proposal is still at pre-application ^[2], which means the objection has formed against a height and a footprint rather than against a scheme with conditions, mitigation and a noise envelope attached. That ordering matters, because a position taken in the absence of a document has nothing in it to negotiate away. The DatacenterDynamics column describing the site argues that engagement should begin before positions become entrenched, and that standardised consultation will not address concerns that are inherently local ^[12]. At 10,000 signatures, the entrenching has happened.

Set the petition against the capital. Two billion pounds spread across 10,000 signatures is roughly £200,000 of proposed investment per person who bothered to sign ^[14]. That is not an argument for capitulation, it is a measure of how asymmetrical the leverage is: adding a signature is close to free, and the schedule it can move is not.

The three cases in the source are aimed at different levers, which is worth noticing if you are drafting a submission. Outside Melbourne, objectors went after operating parameters, specifically prolonged construction noise and a plan to increase the number of backup diesel generators on a hyperscale campus ^[5]. In Inzai, near Tokyo, residents went around the developer entirely and sued the private inspection company that approved a 52-metre building next door, asking for the permit to be nullified ^[6]. One contests how the thing runs. The other contests whether the approval was validly given at all.

The cost of this is described in the piece but never priced. It lists delayed approvals, legal challenges and prolonged planning disputes as sources of investor uncertainty, higher project cost and slower delivery ^[7], and notes that power, land and grid capacity are constrained before opposition enters the picture ^[8]. No months, no percentages, no worked example. That absence is the reason siting risk keeps sitting in the communications budget rather than the capacity plan: nobody has to defend a figure for it in front of an investment committee.

The uncomfortable part is structural. The attributes that make a site score well, open land, available power and a quiet setting, are precisely the attributes the neighbours are defending ^[9]. The site selection model therefore contains the reason the site may not be deliverable, and it contains it as a positive.

Meanwhile the pipeline grows. Analysts quoted in the piece expect global demand for data centre capacity to keep rising sharply for the rest of the decade on the back of AI workloads, requiring facilities at a scale the industry has not attempted ^[10], and governments increasingly treat digital infrastructure as a strategic asset rather than another real estate class ^[11]. More proposals, in more places that were chosen for being quiet. Note also where the argument for early transparency is coming from: an opinion piece in the industry's own trade press ^[4]. That is usually the sign that somebody has already paid for the alternative.