The Week in Energy — 22 August 2026

Gas is ‘insurance’, says the government. The premium isn’t priced.

The week Whitehall conceded the thing it’s legislating to abolish, and tried to answer adequacy in gigawatts.

Kept on board, rarely launched, never optional.

There is a trick to reading an energy document — “find the unit”. This week supplied a master class in it. A storage plan stated in gigawatts is telling you how fast it can pour, not how much is in the jug. A reactor business case stated in jobs is telling you what it employs, not what it costs. A consultation that promises a ‘commercial model’ is telling you there will be a bill, not what the bill will be. Seven days, six documents, and in every one the number that would settle the argument is the number that isn’t there.

START WITH the frankest. On Tuesday DESNZ published its interim response to the ‘Gas system in transition’ consultation. The department’s own words are the finding. Gas provides ‘baseload supply’ and ‘insurance for low probability, high stress events’; there is a ‘continuing need for peak-day capacity’; the three priorities are to maintain gas infrastructure, keep it resilient, and develop ‘a commercial model supporting effective operation’ — a way of paying for plant and pipes that, by design, will run rarely. This is the same government legislating to ‘break the link’ between gas and electricity prices, conceding in a parallel document that gas is the policy the system can’t be run without, and that a new payment stream will be needed to keep it there. The premium is not priced. It will be. The firming cost the ¢/kWh comparison leaves out is the machinery The Energy Trap is built around.

The premium is being paid daily, meanwhile, in the season we’re told renewables dominate. NESO’s own daily fuel-mix posts had gas supplying between 31% and 45% of British electricity across four consecutive mid-August days, and on one of them wind managed 4.6%. Day-ahead gas sat around 150p a therm and baseload power at £110–120 a megawatt-hour, the highest in a year, with European storage at its lowest for the date since 2011 and French reactors derated by warm rivers. David Turver put the long-run bill on this: industrial electricity 63% above the IEA median, industrial energy use down 45% since 2004, and output per head growing 0.4% a year. A 45% fall in industrial energy use is being touted as efficiency. On his reading it is deindustrialisation, and the marginal price is still set by gas in August.

Then the storage answer, which arrived three times and never once in the right unit. Carbon Brief’s explainer on long-duration storage gives the plan: 2.8 GW today, 4–6 GW by 2030, 13–17 GW by 2050, a first round of sixteen cap-and-floor projects totalling 7.6 GW at durations of 8 to 32 hours. DESNZ followed with a £28m ‘Ultra-LDES’ challenge for storage that can discharge for over 100 hours. And on Friday the minister consented Beacon Fen], 400 MW of Lincolnshire solar with a battery of ‘up to 600 MVA’ — a power rating, no duration given. Adequacy is an energy question: the terawatt-hours you can dispatch through a still week in January. The Royal Society estimates we’ll need around 100 TWh to survive the UK’s inter-year wind variation. Sixteen gigawatts at thirty-two hours is about half a terawatt-hour. Every document states the pour rate. None states the jug.

The week’s loudest argument was two spreadsheets shouting past each other. Onward’s Firm Foundations claims Britain could save over £320bn by swapping net-zero policy for gas and nuclear; Carbon Brief’s factcheck lists ten flaws, and some are fair — gas plant at £650 a kilowatt when current American projects are running £1,467–2,054, a network ‘saving’ that cuts transmission spend 86% while still plugging in 32 million EVs. But look at the rebuttal’s own figure for the cost of integrating renewables: a mainstream range of £26 to £75 a megawatt-hour. A spread of three to one is not a rebuttal; it’s an admission that nobody knows the number. And the quantity that decides the question — the energy deliverable through a still winter week — appears in neither report. When both sides of a cost war leave the same cell blank, the cell is the story.

The best news of the week was nuclear, and it deserves a number too. Great British Energy – Nuclear published its Full Business Case for three Rolls-Royce small modular reactors at Wylfa — about 1.5 GW of firm, dispatchable power from the mid-2030s, backed by up to £599m of National Wealth Fund finance for design. That is the right kind of capacity, the kind the fuel-mix figures above are crying out for, and it’s a commitment to be glad of. But the case runs the full five Green Book chapters and contains no strike price, no pounds per megawatt-hour and no benefit-cost ratio: value for money is ‘positive marginal value … with adjustments for uncertainty, risk and optimism bias’, inside a £2.6bn envelope the document admits the contract will outrun. Nuclear wins the adequacy argument on the numbers. But we should be allowed to see them.

Ed Conway added the physics the ‘just import it’ position skips. British appliances were built for lean, dry southern-North-Sea gas, so imported LNG has to be nitrogen-blended to meet UK spec — ‘the single most energy-intensive process’ at the Isle of Grain — on top of the energy already spent chilling it to −162°C for the voyage. Every barrel not produced here returns as LNG at a higher energy cost and a higher lifecycle carbon count that the domestic ledger doesn’t show, which is why Norway, peaking five years after us, has now produced more. Even Dale Vince, who called Rosebank ‘a betrayal’, backed both fields on Newsnight this week. The catch is the price controls he wants attached.

Seven days, six documents, one blank. The unit that flatters is always supplied; the unit that decides is left for someone else to find. Subscribe, and the finding comes to you every Saturday.

---

The Energy Trap: Why the Renewable Energy Transition Can’t Work — And What Can is published by Swift Press in September.