Our NL solar forecast is 35-40% more accurate than the national benchmark in the first few hours ahead, and 15-20% more accurate at day-ahead.

Those are the headline results from our first few months of benchmarking against Ned NL, the Dutch national energy dashboard. This post covers what we built, what we measured it against, and why the benchmark question matters so much in the Netherlands.

The forecast brings the approach behind our GB solar forecast to the Dutch grid. In Great Britain, it helped NESO halve its solar forecasting error, with estimated savings of around £30 million a year.

The Netherlands is one of Europe's most solar-dense power systems. Installed capacity reached 29.4 GW at the end of 2025, and solar supplied 19.7% of Dutch electricity production that year. In summer months, solar accounts for close to 40% of generation.

What sets the Dutch market apart is how decentralised that capacity is:

That growth has created real pressure on the grid. Congestion is widespread, and negative prices are now close to a daily occurrence in summer. The Netherlands recorded 581 negative-price hours in 2025, up from 458 in 2024 and the most in Europe. 2026 is on track to exceed that.

Behaviour is shifting in response. Households now pay more to export, which strengthens the case for home batteries. Some owners are angling panels east or west to catch morning and evening hours, when prices are not negative. Each of these changes alters the shape of the solar curve that traders and grid operators need to anticipate. You can't stop negative prices, but a better forecast lets you prepare for them and protect your portfolio.

In the Netherlands, the hardest question isn't how to forecast solar. It's what to measure the forecast against.

Many market participants would expect the official outturn on the ENTSO-E Transparency Platform to be the source of truth. For Dutch solar, it isn't. That data only covers the small number of solar farms connected directly to the transmission system. When we compared ten years of ENTSO-E data with Ned NL, ENTSO-E showed peaks of around 600 MW, while Ned NL showed close to 20 GW. The rooftop capacity that makes up most of the Dutch fleet is simply invisible in the transmission data.

That gap makes ENTSO-E very hard to use for fundamental modelling or power-flow analysis. So we used Ned NL (Nationaal Energie Dashboard) as our benchmark comparison instead. Ned NL is a joint initiative involving grid operator TenneT and gas network operator Gasunie, among others. It combines capacity statistics from CBS, the Dutch statistics office, drawn from subsidy registrations, feed-in data and inverter manufacturers, to estimate historical and current generation. Our conversations with traders and partners pointed to Ned NL as the most credible national estimate available.

Ned NL also sets a clear benchmark. It publishes its own forecast, so we can compare directly like for like. We know clients often have their own curtailment assumptions, so we publish both a curtailed and an uncurtailed version of the forecast.

After 3 months of live benchmarking this summer our forecast outperformed Ned NL's at every horizon we tested.

Intraday, first few hours ahead -> 30-40% better

Day-ahead -> 15-20% better

The largest gains come in the short term, where live satellite data has the most to add. But the advantage holds all the way out to day-ahead, which matters for traders positioning ahead of the auction. We’ve also had strong early feedback from trial users.

For balanced responsible parties, traders and portfolio managers in the Dutch market, a 30-40% reduction in short-term error means tighter imbalance positions and earlier warnings of negative-price periods. A 15-20% gain at day-ahead means better-informed bids into the auction.

The Netherlands is the start of our expansion into European solar, with Germany and France in development.

We offer a one-year backtest to anyone evaluating the forecast, and a one-month trial with access to the user interface and API.