emsvpp
AI AnalysisClassification & Focus Areas
"Renewable power forecasting combining physics-constrained AI models and automated agent — reduces imbalance costs for energy traders across European markets."
Domain Affinity
Technology Breakdown
Offerings Products & Services
EPT-2
Jua's flagship foundation weather model providing high-resolution global atmospheric forecasts with state-of-the-art accuracy, trained on historical and physical simulation data.
Power Forecast
An end-to-end machine learning model forecasting actual renewable energy generation in megawatts, trained directly on real generation data for select European countries.
EPT-1.5 Family
Jua's previous generation global weather models, offering reliable deterministic forecasts with early dissemination options for time-sensitive applications.
Athena
An AI agent for energy trading desks that leverages the EPT-2 world model to answer natural language queries, generate analysis, and simulate market scenarios by integrating weather, generation, and market data.
Case StudiesReferences & Success Stories
AI Demand Forecasting for Utilities
Jua for Energy applies its foundation model EPT-2 and AI agent Athena to utility demand forecasting, serving customers like EDF, EnBW, and Axpo. The solution improves forecast accuracy by 4 percentage points, translating to €1.5–3 million per GW annual savings and enhancing grid stability with 24x daily updates.
AI Solar Power Forecasting: Physics-Based Models Beat ECMWF
Jua's physics-based EPT-2 model outperforms ECMWF on surface solar radiation, enabling solar traders to save approximately €3 million per GW annually through reduced imbalance penalties. The platform, combined with the Athena AI agent, provides automated briefings and faster decision-making for solar portfolio optimization.
AI Energy Price Prediction: Physics-Grounded Forecasting
Jua's EPT-2 physics-grounded AI model enhances energy price prediction by delivering more frequent and accurate weather forecasts, crucial as AI data center demand surges. A 4% accuracy gain can save €1.5–3 million per GW, with the platform integrating over 25 models and automated analysis via Athena.
Athena is Live
Athena, an AI agent for energy traders, monitors weather models, price signals, and portfolio positions, reducing manual research time by 80% and accelerating signal-to-decision by 10x. Deployed across utilities and trading firms covering over 50 GW of renewable assets, it delivers morning briefings and scenario analysis in minutes.
The EPT-2 Family: Towards simulating the future
The EPT-2 family of foundation models achieves state-of-the-art accuracy in atmospheric prediction, outperforming ECMWF and other AI models on wind, temperature, and solar radiation. Deployed on the Jua Earth Intelligence Platform and used by energy traders and utilities, it provides forecasts at any lead time with rapid refresh capabilities.