How much do we know about modelling in the energy industry? Let’s explore… Could you compete with your laptop? Designed for discussion…
Picture a very specific moment.
It’s a Tuesday in mid-February, ten years from now. A stalled storm front over the North Sea has collapsed wind output across Germany and the UK. At the same time, a cold snap in France pushes heating demand to record highs. Millions of commuters plug in their EVs just as daylight fades.
Now imagine being asked to price that hour of electricity — not in real time, but ten years earlier.
That, in essence, is the problem sitting underneath today’s energy transition. It is not forecasting in the casual sense. It is a continent-spanning optimisation problem with physics, weather, human behaviour, and policy layered on top of one another. And yet, hundreds of billions of euros in infrastructure investment — wind farms, interconnectors, batteries — depend on someone getting that problem approximately right.
Get it wrong, and the result isn’t just a write-down. It’s blackouts, stranded assets, or billion-euro batteries sitting idle because a transmission constraint in the Alps only binds on cold Tuesdays in February.
This post looks at the invisible architecture behind those decisions: the modelling stack that determines what is considered “bankable” in the energy transition — and why that architecture is beginning to change.
Bankability as Software
If you are a bank deciding whether to finance a €500 million solar park in Spain, you are almost certainly not starting with a spreadsheet. You are starting with the output of a power-system model that has already earned institutional trust.
