We develop AI machine-learning tools for energy markets. Our models are trained over vast amounts of data to forecast rolling 168 hours (prompt week). Amongst the input features are highly detailed weather, capacity, demand-drivers, seasonality and other features to produce hourly electricity market forecasts of:
All our production models currently apply to the German power market.
Our forecasts beat top in class back-testing results with R2 of up to 98% and Brier scores of 0.03 and lower.
Models are available to premium subscribers via a feature-rich web-based interface or we can offer more direct integration into your operational system.
In development are additional models to forecast:
Machine Learning models projecting 168-hours of power spot prices, national RES production, and residual load to support near-term trading and asset optimisation.
Long-term forecasts for renewable energy capacities and capture prices to support investment analysis, project valuation and strategic planning.
Benchmark your asset or investment opportunity against historic capture price and revenue data for RES and BESS.
Post and browse teasers relating to development or operating assets. See who is doing what in our market news section or connect with other members.
