Site-specific solar & power forecasting
As solar’s share of the mix grows, operators need accurate forecasts across every horizon, from 15-minute intraday updates to day-ahead planning.
NEXT fuses numerical weather prediction, satellite imagery and on-site measurements, and learns the behavior of each plant. Measured on a European IPP’s own production data, NEXT cut forecast error by 33% at 15 minutes, the settlement horizon on most markets, and by 46% at 5 minutes, against the incumbent satellite-based provider. NEXT was also selected at the world’s largest PV plant after a live trial run through the 2025 monsoon, the hardest season of the year for solar forecasting in India.
- Next
Who Is NEXT For?
Independent Solar Power Producers
Optimize your operation strategy with or without storage by maximizing revenues and minimizing imbalance charges
- Reducing forecast deviation penalties
Minimizes penalties imposed by grid operators in the event of discrepancies between forecasted and actual production. - Maximization of revenue
Optimize your strategy between spot market trading, power purchase agreements, ancillary services and battery storage - Efficient storage management
Improves battery management to minimize storage degradation, extend battery life and reduce replacement costs.
Aggregators & Traders
Maximize revenues from your photovoltaic assets with precise, reliable forecasts throughout the day and the day ahead.
- Maximized revenues
Take advantage of price differences between markets and optimize portfolio balancing. - Reduced penalties
Reduce the cost of penalties for forecast inaccuracies. - Enhanced risk assessment
Understand the variability of solar production through probabilistic forecasts, essential for balancing supply and demand
Three pillars behind NEXT's performance
Multi-source data integration
NWP + satellite + on-site SCADA, fused through an adaptive data-driven framework. Captures both regional weather data and local plant-scale variability.
Adaptive machine learning
Learning algorithms to forecast solar energy dynamically reweight data sources per forecast horizon and per site, adapting to microclimate, grid regime, and seasonality.
Deterministic & probabilistic outputs
Single-value forecasts for scheduling, plus full P5–P95 quantiles for risk-aware trading, BESS dispatch, and balancing strategies.
- Next
Intraday and day-ahead forecast
At CalibSun, we provide intraday and day-ahead forecasting for large-scale solar players. We use a unique hybrid approach that combines satellite imagery, weather models, and real-time onsite data to recalibrate our forecasts to local conditions.
This approach achieves best-in-class accuracy, with a 46% improvement at 5 minutes and 33% at 15 minutes that translates into tangible financial gains.
Consistent forecast accuracy across all horizons
NEXT combines dynamically NWP models, geostationary satellite imagery, and real time ground measurements to capture both large-scale atmospheric dynamics and the sub-kilometer variability that drives actual power output. This multi-source integration ensures consistent forecast accuracy across every forecast horizon, from intra-hour schedule to day-ahead operational planning.
Data confidentiality guaranteed : Each customer retains full ownership of their data, which is used solely for their benefit.
Probabilistic (P5 to P95) and deterministic (P50)
NEXT generates probabilistic forecasts natively thanks to on-site data. Those insights allow you to build a variety of scenarios for your future production. The probabilistic approach enables power plant operators, aggregators and traders to better manage the financial risks associated with fluctuations in energy production.
Calibrated to your plant, recalibrated in real time
Every photovoltaic power plant has a unique relationship between solar irradiance and power output. NEXT exploits your historical data to learn this signature, and continuously recalibrates as new measurements flow in. This temporal adaptation maintains peak forecast accuracy over years of operation, across seasonal cycles, plant aging, and evolving weather conditions.
Available globally without hardware installation
NEXT is available worldwide. It uses several weather models (GFS, Arome, Arpege) and different satellites to ensure worldwide coverage (MSG, IOD, GFS GOES, Himawari…). With no hardware installation required, NEXT works by connecting to your on-site sensor system (SCADA) and takes into account relevant data to refine predictions.
Why make the switch to NEXT
- With NEXT
- Site-specific approach with Hybrid NWP + satellite + on-site data, calibrated to your site
- Works with your existing sensors on site
- Real time recalibration with site behavior
- Full probabilistic envelope + deterministic
- No lag and missed forecasts with real time datasets
- Without NEXT
- Generic forecast at 9–25 km resolution, blind to plant-scale variability
- Extra-cost for integrating on-site data
- No plant-specific behavior integration
- Deterministic outputs only
- Data lag from retrieval and processing leading to +20 minutes update frequency
The 15-minute accuracy we get from in-situ data is surprising compared with other models. For BESS operation, that is exactly what we need.
Mathieu DebonnetCEO, TSE, 2026
The onboarding process
From your data to actionable forecasts
- Cases
Our use cases
Built on Scientific rigour
NEXT is developed by CalibSun and backed by scientific research with Mines Paris – PSL, a key contributor to Task 16 of the IEA PVPS, Mines Paris – PSL, inheriting deep expertise in solar radiation modeling, resource assessment, and numerical weather prediction. Peer-reviewed research and continuous R&D published in scientific papers, ensure NEXT keeps pace with current solar forecasting science.
Need more information?
Get in touch!
- Questions
Frequently asked questions on solar forecast for PV
What is the role of solar forecasting in grid stability?
As photovoltaic capacity grows, grid operators need to anticipate solar power generation variability to keep supply and demand balanced. Accurate solar forecasting reduces reserve activation costs, limits curtailments, and strengthens the reliability of solar energy within modern power systems.
What's the difference between short-term forecasting and day-ahead?
Short term forecasting covers 15 minutes to a few hours ahead, relying on real time satellite and on-site data to capture rapid cloud cover changes, ideal for intraday trading and balancing. Day-ahead forecasting extends 24 to 48 hours, using NWP models to support market bidding and scheduling.
How does NEXT handle cloud cover and rapid weather changes?
NEXT combines satellite imagery, NWP models, and on-site sensors, dynamically reweighted by machine learning to track cloud motion and irradiance ramps. For sub-30-minute horizons, optional sky imagers further sharpen nowcasting under fast-changing weather.
What is solar energy forecasting?
Solar energy forecasting predicts future solar power generation from photovoltaic plants, based on weather inputs, historical data, and plant-specific characteristics. It helps operators optimize operations, secure trading revenues, and contribute to a stable, decarbonized grid.
What are the main challenges in solar forecasting?
The primary challenges in solar forecasting include the high variability of solar irradiance due to cloud cover and atmospheric conditions, the limitations of single-source models, and uncertainty in predictions across diverse geographic locations. NEXT addresses these forecasting difficulties by fusing multiple data sources, NWP models, satellite-based data, and on-site historical data, through an adaptive machine learning engine that continuously recalibrates its forecasting models to maintain peak accuracy, regardless of location or weather conditions.
