Wind · Nordics25%
improved day-ahead wind power forecasts
A pilot forecasting wind parks in the Nordics - the result of hylosense winning Fortum's Spark Innovation Challenge.
Agentic, autonomous forecasting
of every energy asset.
AI that forecasts the power output and weather of each energy asset you own, saving utilities millions in imbalance costs and lost revenue.
Trusted by generation, trading and grid teams at utilities worldwide
Utility from Austria
Utility from Greece
Utility from Ireland
Utility from Italy
Utility from Portugal
Utility from Turkey
Utility from Hong Kong
Utility from Greece
Utility from Ireland
Utility from Italy
Utility from Portugal
Utility from Turkey
Utility from Hong Kong
Average models turn forecast errors into a P&L line.
Every megawatt you get wrong is paid for somewhere: in penalties, in bids, in re-trading and in reserves. Average models, built for everyone, make the same mistakes on every asset.
-
Imbalance penaltiesDeviation × imbalance price
-
Mispriced bids & re-tradingDay-ahead bids, then late intraday fixes at worse prices
-
Excess reservesCapacity held for uncertainty
-
Missed revenueOutput curtailed or undeclared
The agentic forecaster working 24/7 on your assets.
One agentic forecaster for each asset, trained on its own data. The one model you can set up and actually forget.
-
Reads everything that moves your output
SCADA, terrain, global weather models and the microclimate around your exact site, catching details other models miss.
-
Trains a model on your asset alone
No regional averages. The agent is calibrated to how your own turbines, panels and units actually behave.
-
Publishes power and weather, hourly
Day-ahead and intraday forecasts delivered into your existing systems via API or email, no changes in your workflow.
-
Catches its own misses and self-corrects
The agent watches every new reading, flags drift and recalibrates before un-seen errors reach your P&L.
Fleet live monitoringhylosense agents monitor each asset and recalibrate automatically
8 agents running
Wind park
42 turbines · 126 MW
AccuracyDrift detectedRecalibrating…Accuracy recovered94.2%
Solar farm
Tracker PV · 48 MWp
AccuracyDrift detectedRecalibrating…Accuracy recovered95.8%
Hydro plant
Run-of-river · 80 MW
AccuracyDrift detectedRecalibrating…Accuracy recovered96.1%
Thermal unit
CCGT · 420 MW
AccuracyDrift detectedRecalibrating…Accuracy recovered97.9%
Geothermal
Binary cycle · 45 MW
AccuracyDrift detectedRecalibrating…Accuracy recovered98.2%
BESS
60 MW · 120 MWh
AccuracyDrift detectedRecalibrating…Accuracy recovered92.7%
Substation SS 04
110 kV · P & Q
AccuracyDrift detectedRecalibrating…Accuracy recovered95.4%
Trading portfolio
Day-ahead price & volume
AccuracyDrift detectedRecalibrating…Accuracy recovered93.1%
Real results, measured on real assets.
Wind · NordicsThermal · UK · DE64%
improved day-ahead generation forecasts
Across gas plants in the UK and Germany, cutting imbalance exposure and sharpening trading positions.
TSO · N. Macedonia93%
day-ahead substation active power forecast accuracy
Forecasting active and reactive power flow through substations, especially ones with ever-changing solar connections.
A global model averages.
An agent learns your asset.
Industry defaultGlobal model
- One model for everyone
- Regional grid in kilometres
- Drifts quietly until someone notices
- A weather feed you convert yourself
- Months of integration work
- Retrained once a year, if ever
hylosenseAgentic forecaster per asset
- One model tailored to each asset
- Calibrated to your exact site
- Catches misses and self-corrects
- Agentic forecasting of power output
- Live in a day, via API or email
- Learns from every new reading
Built by people who have shipped environmental data to millions.
Over 20 years of combined experience analyzing environmental data. Our previous startup, AirCare, built an air-pollution app used by millions around the world. The founders have their own TED talk, and awards from the United Nations and the President of North Macedonia.
- Climate-tech and weather data expert with over 10 years of domain experience.
- Ex-Booking.com technical team manager, leading multiple user-facing teams in Amsterdam.
- Featured by Forbes, NY Times, WIRED and BBC. Center stage at Google's Flutter conference, TEDx speaker.
- Climate-tech and weather data expert with over 5 years of domain experience.
- Enterprise software engineer at WS Audiology, building enterprise and big-data systems.
- Led North Macedonia's first accreditation of informal tech education and mentored 500+ aspiring engineers.
Backed by
Featured by
Forbes / BBC / The Guardian / NY Times / Washington Post / WIRED / TED / HEAPS Japan / The Verge / La Repubblica
Try hylosense agents on your assets
We work with generation, trading and grid teams at major utilities worldwide. Book a demo to see how hylosense plugs into your existing systems and workflows.