The grid changed.
The models didn't.

Britain's electricity system was designed around a few large, predictable power stations. It now runs on thousands of wind farms, solar arrays, batteries and flexible loads that shift by the minute.

Most of the tools used to plan and balance that system still work at yesterday's resolution. The gap shows up as curtailed renewables, unnecessary constraint payments and generation that didn't need to be burned.

Higher resolution,
end to end

We build models that treat the network as it really behaves: weather-driven generation, physical transmission limits, storage that charges and discharges on price, and demand that responds rather than simply arriving.

That means combining live network data with generation forecasts and settlement histories, then resolving them at the timescales decisions are actually made on — from day-ahead planning down to within-hour balancing.

The output isn't another dashboard. It's a forecast operators, traders and asset owners can act on, with the uncertainty stated honestly rather than hidden behind a single number.

What we work on

Three problems, one underlying model of the system.

Forecasting

Generation and demand predicted at the resolution the system is balanced on, with confidence intervals that hold up under real weather.

Constraints

Where the network physically limits what can flow, why it binds, and what it costs when renewable output is turned down to respect it.

Flexibility

How storage, interconnectors and responsive demand can absorb the difference — and what that is worth across a day, a season, a decade.