Make It a Contract, Not a Gift
Guest writer Douglas Guy replies to Craig’s recent article on Zonal Pricing and our Daily Briefing on Data Centre politics by looking at the demand side of the equation and asks if data centres could be supported if they could meet conditions around ownership and environmental impact.
Common Weal explained why we pay wind turbines to stop. Then told us to block the one thing that could use it. Both can't be right. The one thing is a 600MW data-centre campaign at Auchtertool in Fife, the largest parcel of new demand anyone has offered Scotland, filed under things to avoid. This is an argument for a conditional yes: the data centre is worth having, but only on terms that make it serve the grid instead of merely sitting on it.
Start with the physics: the politics keeps skipping it. A wind farm is generation; a data centre is load. Scotland's grid problem is not too little generation but too much, stranded behind a wire too thin. The Scotland-England boundary, B6, carries about 4.5GW and was constrained for roughly a third of 2025; the subsea reinforcement that relieves it is not energised before 2029. Until then, every megawatt of demand sited north of that bottleneck is a megawatt of Scottish wind that need not fight its way south or be switched off. Auchtertool is in Fife. It is on the right side of the wall. To file "corporate warehouses" alongside corporate wind farms as one imposition is to oppose the disease and the cure at once: one makes the surplus, the other could drink it when it spills.
Follow the money and it does not go where the slogan says. The 2024/25 balancing bill came to £2.7bn, the rise driven almost entirely by constraints; thermal constraint costs alone hit £1.7bn, up 64 per cent in a year. "We pay wind to switch off" captures the minority share. The larger cost is paying southern gas plants to fire up in place of the power the wire could not carry. The reinvestment fix, keeping constraint payments but spending rather than banking them, aims at the cheaper share of the bill and leaves the costlier one standing. A flexible load in Fife goes after the expensive part directly: in the full-wire hours, use the wind where it lands, and pay neither the turbine to stop nor a gas plant in Yorkshire to start.
So far this sounds like a brochure for hyperscale AI. It is not, and the conditions are the whole argument. A careful NBER working paper is more honest than any developer's slide: temporal flexibility always lowers system cost, but its effect on emissions depends on the grid it sits in. On a wind-rich grid a flexible load chases the surplus and cuts carbon; on a fossil-heavy one the same flexibility merely smooths the baseload plant's day and raises it. Flexibility is not green by nature. It is green by location. Which is precisely the case for Fife, and precisely why a Fife consent has to bite.
And it has to bite on the right things: the load can move. The work that wants the wind need not sit in Fife: a hyperscaler can lift a deferrable job down a cable to another of its sites and run it on Norwegian hydro that does not care what the North Sea is doing, though using it costs money and a single site run flat out is cheaper. So pin the conditions to what cannot follow the data across the water: the grid connection, the heat the thing pours out, the public stake in the shed itself, and the law it answers to. Servers on Fife soil still sit under a foreign master's jurisdiction; the US CLOUD Act reaches their data wherever the field stands. Whether Scotland should own the compute, not just the ground beneath it, is a fight bigger than one 600MW shed can settle. When the developer says the scheme will have "zero impact" because it draws on private wires and power-purchase agreements, that is over-claimed: a continuous 600MW load cannot have zero impact, and such an agreement is a contract, not a physical promise that the servers stop when the wind drops. Real interruptibility is the one thing the application does not yet contain.
The law for this is already on the shelf, unused. This is a Permission in Principle; planning authorities can attach conditions and time limits, and a consent left unbuilt lapses. That lapse is the point. Britain spent 2025 learning why use-it-or-lose-it matters, as the grid connection queue swelled past 720GW with speculative placeholders, and rewrote the connections regime around it. Planning should carry the same teeth. Make the consent a contract, not a gift: time-limited, interruptible, locally beneficial, and revocable. The one who means it will sign; the one who is landbanking will walk, and good riddance. A moratorium, the campaigners' remedy, is only a slower way of saying no: diagnose the vacuum, then fill it.
“If the public should own a stake in the generation, it should own one in the load that makes the generation pay.”
Common Weal already hold the right tool. If the public should own a stake in the generation, it should own one in the load that makes the generation pay. That is how a country captures value rather than hosting a shed.
All of this governs the hours when there is too much power. The harder question is the hours when there is none. On the still, frozen evening when the wind does not blow, a flexible load, however well-behaved, has nothing to flex against. A dunkelflaute is not a peak that clears by nine; it is a high-pressure system parked over the North Sea for a week, and no clever scheduling rides that out. With Torness closing, Scotland's firm floor on that night is Peterhead burning gas. On this Common Weal are simply right, and ahead of the briefing: the firm low-carbon power and long-duration storage the transition needs is exactly the concrete a privatised, nationally-priced market will never pour, and only a public mandate ever has. Even that has limits: the cap-and-floor backstop buys pumped storage that smooths the surplus, not firm power for the deficit, and the one firm source big enough is the nuclear the country has ruled out. That deserves its own essay; but the flexible-load case should stop pretending the dark is someone else's problem.
So the position is conditional, and the condition is the point: yes, if it is sited where the wind is stranded, flexible enough to follow it, cooled without draining the burn, owned in part by the people whose grid it leans on, and consented on terms that bite. Refuse it on those terms and the developer should leave. Refuse it on reflex, because the thing is big and corporate and new, and we keep the worst of all worlds: the wind still curtailed, the gas still burning in Yorkshire, and the compute built somewhere that had the wit to say yes. Today's heresy is next year's planning condition, and a dead calm powers no compute.
—
This article is a condensed version of two articles published by the author at airt.scot · the argument in full: Generation, Meet Load (https://airt.scot/essays/generation-meet-load/) and The Windless Dark (https://airt.scot/essays/the-windless-dark/). Free to use, in whole or chopped into little pieces, provided the source is acknowledged (CC BY 4.0).