Machine learning for electricity demand forecasting

  • Investment Strategy,
  • ThemeClean Energy & Transportation
  • LocationHouston, Texas
  • Date of Initial Investment2022
  • Bridges ExecutivesMike D’Aurizio

Thesis

Electricity markets depend on accurate demand and generation forecasts to determine how much power to produce, purchase, and deliver. Rising intermittent renewable generation, climate change-driven weather volatility, and evolving demand factors (e.g., EV charging, rooftop solar) are increasing both the complexity and cost of forecast errors. Traditional, often spreadsheet-based, forecasting methods struggle to capture these dynamic conditions, exposing energy providers to sudden price volatility and undue market risk.

Investment

Amperon provides electricity demand, renewable generation, and price forecasts to power producers, utilities, and energy retailers. It leverages proprietary machine learning models trained on grid, smart meter, and weather data to deliver forecasts that consistently outperform independent system operator benchmarks. By improving forecast accuracy across time horizons and system levels, Amperon enables more efficient renewable integration and reduces exposure to wholesale market volatility that otherwise would drive up costs for consumers.

Outcomes

Amperon’s forecasts mitigate the reliability challenges of weather-dependent power, making it easier for grids to integrate renewable capacity.