Solar Asset Intelligence

Production data arrived from four inverter vendors in four formats, at four different intervals. Yield reports were assembled by hand at the end of each month, and by the time an underperforming string was noticed, it had usually been underperforming for weeks.

The operator needed one platform: live telemetry, financial yield tracking, and production forecasting that the asset managers could defend in front of their investors.

We started in the field, not the IDE. Our co-founder spent a decade operating solar assets, so the data model was designed around how a farm actually degrades: soiling, string faults, inverter clipping, curtailment.

A unified ingestion layer normalises vendor telemetry into a single time-series store. On top of it sit anomaly detection tuned to catch string-level underperformance within hours, and a forecasting model that blends irradiance data with each site’s measured behaviour.

The platform shipped as one piece: ingestion, analytics, alerting, and investor reporting — no glue code between vendor tools.

Asset managers now open one screen in the morning instead of six. Underperformance is flagged the same day, monthly investor reports generate themselves, and forecasts are built on the portfolio’s own measured physics rather than nameplate optimism.