70 Cars, 55,000 hours, one question: Can you forecast when an EV battery will die?

Datakrew, a Singaporean deeptech startup, has completed a year-long study in partnership with Hyundai CRADLE, the automaker's corporate venture arm, and GetGo, Singapore's largest car-sharing platform. The research involved 70 Hyundai EVs, logging over 3.6 billion data points, 101,000 trips, and 55,000 operating hours. The study aimed to forecast the State of Health (SOH) and Remaining Useful Life (RUL) of EV batteries, which typically degrade slowly and unevenly due to various factors like driving habits, charging methods, and environmental conditions.

Datakrew's IoT device, ITUS Max, and its analytics platform, OXRED MyFleet, gathered data on battery and telemetry parameters every ten seconds for a full year. The machine learning system demonstrated a prediction error band of three percent at various forecast horizons, including long-range forecasts of up to two years. The company's architecture supports multiple OEMs and vehicle models, with ongoing expansion into the UK, Europe, and Mexico, as well as potential insurance applications.