Session chair: Farzaneh Pourahmadi – DTU

AI is increasingly used in electricity markets for applications such as forecasting, bidding, multi-market trading, and demand flexibility. Yet, most existing approaches remain task-specific and can struggle to generalize across changing market conditions, market designs, products, and participants. This session will bring together recent developments in AI for power markets and discuss where foundation models may offer genuine advantages over classical AI approaches, where they may not, and what methodological and practical advances are needed for their reliable deployment. Ultimately, the session will explore whether a “MarketFM” could emerge as a useful foundation for future market forecasting and decision-making.

Foundation Model for Electricity Price Forecasting

Speaker: Runyao Yu – TU Delft and AIT

Electricity price forecasting in Europe presents unique challenges due to increasing renewable generation variability, market integration, and the continent’s physically interconnected power system. While recent advances in foundation models have led to substantial improvements in general time series forecasting, most existing approaches do not incorporate prior graph knowledge from the transmission topology, which can limit their ability to exploit meaningful cross-region dependencies in interconnected power systems, motivating a domain-specific foundation model. In this presentation, we show how to design and use the pretrained foundation model for probabilistic electricity price forecasting.

Foundation models for electricity price forecasting and battery arbitrage: Can they replace market-specific forecasting models?

Speaker: Rafał Weron – Wrocław Tech

Foundation models promise accurate forecasts with little or no task-specific training, but whether they can replace models designed specifically for electricity price forecasting remains unclear. We compare nine variants from five foundation model families, evaluated in zero-shot mode, with two state-of-the-art electricity price forecasting benchmarks in Germany, Poland, and Spain over 2021-2025. Their performance is assessed in terms of point and probabilistic forecasting accuracy, as well as economic value in battery energy storage arbitrage. Only the TabPFN models consistently and significantly outperform the benchmarks across all three markets and all statistical measures. However, this statistical dominance does not translate directly into economic dominance: TabPFN performs best under unlimited bids and riskier quantile-based strategies, whereas the Distributional Deep Neural Network benchmark is more profitable when risk tolerance is lower. Thus, foundation models cannot universally replace market-specific models, and their value depends on both model architecture and the decision problem.

Trading blind to the grid, why price forecasts aren’t enough

Anna Varbella – kyra/ ETHZ

AI in power markets today is mostly used to forecast prices and load, and models are ranked by statistical accuracy. But a more accurate forecast does not necessarily earn more money. To know what a forecast is worth, you have to measure the value it adds when it drives real dispatch decisions. A second gap is that most of these models ignore the grid itself. Batteries and data centers are optimized with no view of network constraints, and curtailments and redispatch follow. Kyra builds a physics-informed digital twin of the European transmission grid at 60 kV and above. We use it to find nodes with spare capacity for new BESS and data centers. For operating assets, we publish grid-feasible volumes every 15 minutes, giving traders an operating band they can optimize within without running into curtailment.

When and How Should a Power Trader Engage in Arbitrage? Predict, then Contextually Optimize

Yannick Heiser – DTU

Electricity markets increasingly expose stochastic energy generators to
arbitrage opportunities between the day-ahead and balancing markets, driven by widening price spreads. However, opportunistic bidding, deliberately deviating from the
production forecast to exploit anticipated price spreads, carries significant risk, and
existing frameworks rarely offer explainable, risk-aware decision support. We propose
a predict-then-contextual-optimize framework that decomposes the day-ahead bidding
decision into three explicit stages to decide, when to engage in arbitrage, in what direction, and to what extent. A probabilistic binary classifier with confidence thresholds
determines whether the predicted price spread is sufficiently confident to justify an
opportunistic bid. Otherwise, the trader defaults to an arbitrage-free bid equal to the
power forecast. A linear decision policy learned for each class via contextual optimization determines the magnitude of the bid deviation from the power forecast. The
framework accommodates both standalone renewable generation and hybrid power
plants combining renewable generation with other assets, such as an electrolyzer. We
evaluate the framework on a real wind farm in the European bidding zones DK1 and
DE/LU using a rolling-window procedure and compare it against several benchmark
bidding strategies. The results show that the proposed framework increases mean
profit relative to an arbitrage-free benchmark, reaching an improvement of about 7%
for the hybrid power plant in DK1. The largest gains occur when distributional drift
between training and testing windows is low, while the co-located electrolyzer further
increases arbitrage value by providing additional operational flexibility.