Two notices in one week
During the week of 22 June 2026, Britain's National Energy System Operator issued Electricity Margin Notices on the 24th and again on the 26th. A margin notice is the system operator telling the market it wants more generation available than it currently expects to have. It is not a blackout and it is not a crisis, but it is the clearest routine signal any operator gets that the system is tighter than planned. Most industrial and commercial energy buyers who track anything at all track these.
On 17 July, the energy regulator Ofgem announced a formal post-event review of that week. It said it takes "the recent allegations concerning the operation of the electricity system during the period of extreme heat in June, and the associated whistleblowing concerns, extremely seriously". The review is being conducted under Condition C7.5 of the ESO Licence, which covers energy resilience and resilience reporting, and its named oversight includes Jayne Scott, chair of the Audit and Risk Assurance Committee, Siobhan Duffy, shareholder non-executive director, and Akshay Kaul, a director-general at Ofgem.
Read the scope, not the headlines
The list of things Ofgem says it will examine is where the useful information sits. The review covers the weather conditions of that week, the operational decisions taken, compliance with security standards, and the sufficiency of record-keeping. That last item is the one worth stopping on. A regulator does not normally announce that it is checking whether the records of an event are adequate unless there is a live question about whether they are.
Four days earlier, on 13 July, NESO had commissioned the law firm Eversheds Sutherland to run an independent investigation into whistleblowing allegations concerning operational decision making and record-keeping. NESO said staff would be able to speak to the investigators anonymously. The allegations, as reported, include a claim that control room staff were discouraged from creating a paper trail during the late-June heat. NESO has not accepted that characterisation, and the investigation exists precisely because the question is open.
Why this lands on an operator's desk and not just a politician's
Most of the coverage has treated this as a governance row, and on its own terms that is fair enough. But there is a narrower consequence that belongs to anyone who runs a site in Great Britain, and it has nothing to do with who said what to whom.
If you have ever sized backup generation, negotiated an interruptible tariff, chosen a data-centre location, or written a resilience section in a board paper, you almost certainly used the public record of system stress as an input. Margin notices, post-event reports and the operator's own retrospectives are the cheapest available evidence about how often the system runs close to its limits. That evidence is only as good as the record it was drawn from, and the completeness of that record for June 2026 is currently the subject of a regulatory review and an external legal investigation.
This does not mean the published record is wrong. It means it is provisional, which is a different and more manageable statement. The correct response is not alarm. It is to stop treating one particular week's public account as a fixed input in your own modelling until the body that regulates the system operator has said whether it holds.
The question these reviews will not answer
There is a limit worth knowing in advance. Both processes are pointed at conduct and documentation: what was decided, what was written down, whether the rules were followed. Critics of the terms of reference have argued publicly that they leave out the central question of whether system security was actually at risk that week. Whether or not you accept that criticism, the scope as published is about decisions and records rather than about margins and reserves.
So the realistic outcome is a finding about process integrity. That matters, and a regulator confirming that the record can be trusted is genuinely useful. What will not arrive is a verdict on whether Britain's grid can carry the load being planned for it. Anyone deferring an investment decision until these reviews conclude should be clear that they are waiting for the wrong document.
What to change this week, and what not to
Do not rip up a resilience plan over an unresolved investigation. Do go back to whichever document justified your current backup capacity or your interruptible contract, find the line where it cites system-stress history, and mark it as unverified pending Ofgem's report. That is a five-minute annotation and it stops a provisional number from quietly hardening into an assumption nobody revisits.
Then get one input you own. Your own half-hourly load data, and your own record of voltage events and interruptions at your sites, do not depend on anyone else's minute-taking. Operators who can describe their own exposure from their own meters are not waiting on this review to make a decision, and that is the position worth being in whichever way the findings land.
Read next: Twenty-Seven Ministers Named, No Tech Secretary | Britain Cancels Digital ID, Not Your ID Duty
