June’s heatwave: prices said one thing, the wires said another

By Roger Hey CEng, Managing Director, Smart GB

On Monday 19 October 2026, Ofgem and the National Energy System Operator (NESO), which runs Britain’s electricity system minute by minute, published their first reports on June’s heatwave. The good news is real. Nobody lost power, and the system stayed inside its legal limits.

I’ve now read both reports, and I’m more worried than reassured. For much of that week, the electricity market and the physical network were pulling in different directions. The people in the control room spent their evenings holding the two together by hand.

What happened on 23 June

Some context first. Britain is linked to France, Belgium and the Netherlands by undersea cables called “interconnectors”. Power flows through them towards whichever side is paying more. Most of the time, that helps everyone.

On 23 June, the whole of western Europe was baking. Most electricity is bought and sold the day before it’s used, and on the Monday the prices for Tuesday evening looked much the same here and in France, at around £240 a megawatt hour. On that basis, the market planned for France to send us about 3.5GW that evening.

Then, during Tuesday itself, French prices rocketed as their own heat-hit power stations and soaring demand caught up with them. By late afternoon, France was paying around £1,200, roughly double our price. Traders can keep buying and selling right up to the last hour, and they did. The planned imports from France turned into exports to France. The swing was about 5GW, the output of four or five large gas power stations.

NESO bought power back to protect the system. Other traders, quite legitimately chasing the French price, sold in the opposite direction. The market traders and the control room were working against each other.

The grid has a heartbeat, its frequency, which should sit at 50 hertz. It dips when demand outruns supply. That evening it stayed below NESO’s normal range for long spells and fell to about 49.66. In the end, NESO engineers ordered the Dutch cable to stop exporting and asked France for emergency help. That did the trick, just after 8pm.

A message sent the wrong way

One detail stopped me in my tracks. Earlier that afternoon, NESO wanted to tell its French counterpart, RTE, that Britain couldn’t offer France emergency help for a couple of hours, because of congestion on the transmission network in the south east. According to NESO’s own account to the review, the message went across the wrong way round, describing the opposite direction, from France to Britain.

Half an hour later, RTE withdrew its emergency support to Britain for 7pm to 8pm. That turned out to be the worst hour of the night. The report doesn’t say one caused the other, and I won’t either. But it tells you a lot about how stretched and chaotic things were.

Not much in reserve

Across the week, the report counts about 40 times when frequency left NESO’s normal range. It counts about 600 times when it strayed beyond the starting point NESO assumes when planning for a sudden fault, such as a power station tripping off. That matters. If a fault hits when you’re already off balance, you have less in hand to recover.

And faults happen. At a quarter to two on the morning of Friday 26 June, a large transformer near Liverpool failed and caused a small fire. Losing that circuit cut how much power could move around the network. For a few hours, NESO’s plan for the evening showed a shortfall of over 3,500MW if it stayed out. Thankfully it was back by half past six that morning.

By the afternoon, Britain as a whole had almost enough power. The problem was getting it into London and the south east, where five transmission circuits were out of service. Emergency action stayed on the table into the evening.

That afternoon exposes an assumption buried deep in how we trade electricity. Britain buys and sells power as if the whole country were a single point, with one national price whatever the wires can actually carry. Engineers call this a “copper plate”, an imaginary network with no limits. The real network has very definite limits, and that day they sat between where the power was and where it was needed. In effect, we’ve built the market around physics that doesn’t exist, and left the control room to deal with the physics that does.

Blind spots and nerves

Two more findings bother me. On 24 June, NESO found a gap of around 5,000MW between what generators had published the day before and what they were telling the control room. That’s a big blind spot to carry into a heatwave. The report does not state a cause, but I wonder if it was poor data or an IT system glitch.

Separately, the NESO Board asked the law firm Eversheds Sutherland to investigate whistleblowing allegations about that week. Its brief was narrow. It looked at whether records had been deliberately altered or audit trails not kept, and whether anyone had interfered with operational decisions. It was not asked to judge how much risk the system was carrying. That is for Ofgem’s inquiry. Eversheds found no deliberate intent to amend documents. But it did identify what Ofgem calls a “culture of reputational sensitivity”, including some reluctance among control room staff to issue public warnings that margins were tight. Those warnings exist so the market can respond. If people hesitate to send them, the system loses one of its few early levers.

The question Ofgem is asking

Ofgem now wants to know whether our market can reliably “self-balance” when things get tight, or whether the arrangements need to change. I think June gives a fairly clear answer.

Resilience has become oddly tribal. One camp says it’s all fine. The other predicts doom and stops there. Neither is much use to an engineer watching 5GW reverse direction at teatime. What June shows is more mundane, and more fixable. When prices point one way and the physics points the other, someone has to bridge the gap by hand. And in a Europe-wide heatwave, the tools they reach for, imports and help from our neighbours, are under strain too.

What happens next

NESO’s final report is due by 11 December. Ofgem will say whether the system was secure by 31 January, and publish its lessons by 31 March. I’ll be looking for straight answers on three things: why prices and physical conditions drifted so far apart, why the control room’s data was so far out, and whether people felt able to raise the alarm.

The lights stayed on in June. I’d rather next summer didn’t depend on emergency calls to Paris.

This article was first published on LinkedIn, where you can join the discussion.