Verified against more than 10,000 ground stations with no post-processing, EPT-2 beats ECMWF HRES on 10 m and 100 m wind, 2 m temperature and solar radiation across the full 0–240 hour range. Its ensemble, EPT-2e, beats the 50-member ECMWF ENS mean at virtually every lead time. Our research is peer-reviewed at ICLR and NeurIPS.
The same models serve the trading desk, the control room and the fund.
Athena watches your positions and every weather and market model, and flags what is about to shift on your spreads. EPT-2 Early lands ahead of the forecasts other desks are waiting for.
Athena for energy trading →Wind and solar forecasts turned into power output for your fleet, with actuals nowcasts, 5 km models over Europe, and alerts when the forecast shifts.
Talk to us →Deterministic and ensemble forecasts through an API and Python SDK, with decades of history to backtest against and probabilistic forecasts out to 60 days.
Talk to us →EPT-2 learned the physics of the atmosphere from data. The same base model, lightly fine-tuned, has already handled airfoils and shock waves. We are building a foundation model for the physical world, starting with the markets that need it most.