Currently, energy traders use weather forecasts up to 15 days out and broad seasonal outlooks to buy and sell energy commodities like electricity, oil, and gas, to balance supply and demand and profit from price fluctuations. Temperature changes can affect heating and cooling needs, while wind, solar, and precipitation influence renewable energy generation and hydroelectric power. Accurate weather forecasts can help energy traders anticipate impacts to energy and optimize their trades.

A tech firm called Amperon Holdings Inc. is leveraging artificial intelligence and machine learning to create a range of hourly demand up to seven months in advancethat will be updated upon the daily release of global weather models created by Europe’s biggest forecasting center.

The accuracy of these new projections will depend on the underlying weather forecasts from the European Center for Medium-Range Weather Forecasting (ECMWF), providing forecast conditions every six hours out to seven months. These weather forecasts will be combined with machine learning to come up with hourly demand curves with the help of historical data and solar and wind conditions.

Amperon has already been using machine learning to boost the granularity of ECMWF models and provide hourly temperature forecasts over a seven month period and it has shown early signs of providing accurate forecasts. One of its models successfully predicted a June energy demand spike on the 13-state grid 32 days in advance, as well as impacts of a January 2024 blast of Arctic air that hit Texas 38 days in advance.

Accurate forecasts this far in advance can be beneficial for utilities and power retailers to gain greater visibility into weather conditions and changing consumer behavior to hedge against price shocks. Energy traders may also get a competitive financial advantage.