The AI Stack Map
Electricity is a physical constraint connecting national energy systems and computing infrastructure.
Britain has spent two decades building the most expensive electricity system in the developed world — and still hasn't decarbonised it. The problem was never the technology. It was the market.
In September 2024, Britain switched off its last coal-fired power station. It was greeted as a landmark moment. Politicians gave speeches. Headlines ran warm.
Six months later, the data arrived. In 2025, a record year for British renewables, gas generation increased by 5%. Not fell. Increased. The reason: nuclear output collapsed to its lowest level in half a century, and the gap had to be filled by gas.
This is the central paradox of British energy policy. The country has invested tens of billions in renewables, built the world's largest offshore wind fleet, and genuinely reduced coal to zero. And yet it remains locked in a structural dependency on gas.
The UK's electricity system has been transformed over the past fifteen years. In 2010, coal provided roughly 30% of the nation's power. By 2025, that figure was zero. Wind capacity has grown from under 5 GW to 33 GW. Solar has gone from almost nothing to 21 GW.
But gas remains the single largest dispatchable source of power. Nuclear has been allowed to wither: just 36 TWh in 2025, its lowest output in fifty years.
Biomass — burning wood pellets, mostly imported from North America — has quietly grown to 41 TWh, now generating more electricity than nuclear. Nobody talks about that.
The overall picture is a grid that is genuinely cleaner, but not as clean as it should be given the investment made, and still fundamentally dependent on gas.
Between 2004 and 2019, UK domestic electricity prices roughly doubled — from around 7.5p per kWh to 17p. Green levies now account for 16% of the final price, adding roughly £140 to the average household's annual bill.
Then came the energy crisis. Between autumn 2021 and winter 2022, prices nearly doubled again — from 17p to 34p per kWh. Countries with less gas dependency — France, Norway, Sweden — were far less exposed.
By 2025, prices had fallen back to around 24.5p — still 43% above pre-crisis levels.
In the UK, gas sets the wholesale electricity price 98% of the time — even when wind turbines are generating at full capacity and the marginal cost of that electricity is effectively zero. This is not a law of physics. It is a market design choice. In France, gas sets the price only 7% of the time.
This is the mechanism that creates the paradox. It was designed for a world of dispatchable fossil fuel plants. It was never redesigned for intermittent renewables.
The result: more wind capacity does not automatically mean lower bills. A Common Wealth thinktank analysis found that decoupling electricity from gas pricing could cut household bills by up to £203 per year.
In 2025, Britain spent £1.46 billion paying wind farms to switch off. These are called "constraint payments." In 2025, 13% of all hypothetical wind generation was curtailed. Total grid balancing costs reached £2.7 billion.
The LCOE figures that the government publishes — showing offshore wind at £59/MWh — do not include balancing costs, curtailment payments, or the cost of backup gas capacity. The true cost of an intermittency-dependent grid is substantially higher than headline numbers suggest.
The government's own report shows:
Gas CCGT at £45/MWh at 30% load factor is competitive with onshore wind at £41/MWh before system costs. This is not an argument for burning more gas. It is an argument for nuclear.
Britain is currently paying wind farms £180,000 per hour to switch off. In 2025, the total bill for wind curtailment alone was £1.46 billion. Meanwhile, the government counts the wind capacity as part of its clean energy progress — even the 13% that was never actually used.
The lifecycle story: Onshore wind produces around 11 gCO₂e/kWh. Offshore wind around 14g. Nuclear around 12g. These are all far below gas (490g) and coal (820g).
China controls over 80% of all solar PV manufacturing. A solar panel made in China has a lifecycle carbon footprint of around 65 gCO₂e/kWh — more than five times higher than one made in the EU.
Turbine blades are made from composite materials that cannot be recycled in conventional facilities. The UK's landfill rate for glass fibre composites is 67%. An estimated 800,000 tonnes enter landfill every year globally.
Drax power station burned 15 TWh of wood pellets in 2024, mostly imported from North America. Drax received £893 million in government subsidies in 2023 alone. Biomass generated 41 TWh in 2025 — more than nuclear.
In 1990, the British government privatised the electricity industry. No new nuclear power station was ordered after privatisation. The existing fleet was left to age.
For the next two decades, the "dash for gas" filled the gap. The structural dependency that would cripple Britain thirty years later was being locked in, one gas plant at a time.
Hinkley Point C — the first new nuclear plant in a generation — has risen to an estimated £48 billion in current prices.
Hinkley is, by any measure, an expensive disaster. But the lesson is: nuclear is expensive when you stop building it for thirty years. France built 37 reactors in ten years at one third to one half of the per-unit cost by standardising designs and maintaining a continuous supply chain.
France's nuclear strategy: France generates 70% of its electricity from nuclear. Its grid carbon intensity is 21 gCO₂/kWh — compared to Britain's 126g. Its average industrial electricity price is £69/MWh — compared to Britain's £168/MWh.
Germany's renewable experiment:
Germany's Energiewende invested hundreds of billions in wind and solar and closed all nuclear power stations. The result: the highest household electricity prices in Europe at around 34p/kWh. Its grid carbon intensity remains around 400 gCO₂/kWh — nearly twenty times dirtier than France.
South Korea's mixed strategy: It has pursued nuclear providing close to half its electricity. Its domestic electricity price is around 13p/kWh — roughly half the UK rate. It simply kept building nuclear reactors and did not stop.
On a single day in October 2025, French nuclear output averaged 43 GW. Britain's entire electricity generation from all sources averaged 34 GW. France's nuclear fleet alone generates more electricity than the entire British grid.
| Metric | UK | France | Germany |
|---|---|---|---|
| Grid carbon intensity | 126 gCO₂/kWh | 21 gCO₂/kWh | ~400 gCO₂/kWh |
| Domestic electricity price | ~24.5p/kWh | ~25p/kWh | ~34p/kWh |
| Industrial electricity price | £168/MWh | £69/MWh | ~£200/MWh |
| % of time gas sets price | 98% | 7% | ~50% |
| Nuclear share of generation | 11% (falling) | ~70% | 0% (closed 2023) |
| Renewables share | 47% | ~24% | ~62% |
| Grid balancing costs | £2.7bn/year (rising) | Low (dispatchable) | Very high |
| Net electricity exporter | No (net importer, 10%) | Yes (€5bn/year) | No |
The difference between £69/MWh and £168/MWh is the reason energy-intensive industries have been leaving Britain for decades. It is the reason data centres are choosing to locate in France.
None of this means renewables are a mistake. The argument is not against renewables. The argument is against the idea that renewables alone — without a dispatchable, low-carbon backbone — can deliver a cheap, stable, and clean grid. Every country that has achieved that combination has done it with nuclear at the centre.
The argument has surface appeal but falls apart. Gas prices are structurally volatile — the 2021–22 crisis proved that.
The correct counterfactual is not "renewables vs gas." It is "renewables + nuclear vs renewables + gas." The former delivers a clean, cheap, stable grid.
Britain's electricity is expensive and dirtier than it should be not because of renewables, but because of a market structure that was never redesigned for them — and because the country abandoned nuclear power in 1990. The technology to have cheap, clean, stable electricity has existed since the 1970s. France built it. Britain chose not to.
"Britain has the world's largest offshore wind fleet, pays twice what France pays for electricity, and has a grid six times dirtier. France did it with nuclear reactors built in the 1970s. We called that old technology."
You’ve looked beneath the surface.
Electricity is a physical constraint connecting national energy systems and computing infrastructure.
A connection through “The hidden bottleneck”: Find the physical and institutional constraints beneath apparently limitless systems.
A connection through “The hidden bottleneck”: Find the physical and institutional constraints beneath apparently limitless systems.