We forecast campus energy and clean the air it runs on.
GEST predicts electrical demand and trims the peak before it forms. Sabzpoost is a coating that breaks down airborne pollutants using nothing but the light already in the room.
Loading scene…
Energy loss eliminated
Measured at campus scale
CO₂ avoided per year
About 40 cars off the road
Demand forecast accuracy
LSTM model, 2 kW MAE
VOC degradation
Anatase TiO₂ under ambient light
Less energy than HEPA
No filters, no waste stream
Sensors at 1 Hz
MQTT telemetry mesh
Forecast, reduce, verify
Forecast
An LSTM model reads 300+ sensors at 1 Hz and predicts campus demand two hours out at 95% accuracy.
Reduce
Control acts ahead of the peak, not during it. Lighting drops 5 kW and HVAC drops 8 kW, for 20 kW shaved across the 8am to 2pm window.
Verify
35% less energy lost, 200 tonnes of CO₂ avoided, 90% of VOCs broken down. We publish the measurements, not a marketing round-up.
GEST and Sabzpoost
One system manages a building's electrical load. The other manages its air. Both act on what they measure instead of only logging it.
GEST
Campus energy intelligenceGEST watches more than 300 sensors at 1 Hz and feeds them to an LSTM model that forecasts campus demand two hours out at 95% accuracy. When it sees a peak coming, it acts: it sheds load, banks heat in thermal storage, and draws power from surfaces the building already has.
Sabzpoost
Photocatalytic air coatingSabzpoost is a 20-nanometre anatase TiO₂ coating that runs on the light already in the room. It generates hydroxyl radicals that break VOC molecules apart at the surface: 90% degradation, no filter to change, no moving part to wear out, rated for ten years.
Ask us for the data behind any number.
We work with university facility teams and materials researchers. If you want the measurements behind a figure on this site, email us and we will send them.