PriceFM: Foundation Model for Probabilistic
Electricity Price Forecasting

Runyao Yu1,2,3, Chenhui Gu1, Jochen Stiasny1, Qingsong Wen4,
Wasim Sarwar Dilov3, Lianlian Qi2,5, Jochen L. Cremer1,2
1 Delft University of Technology, 2 Austrian Institute of Technology, 3 Rimac Technology,
4 Squirrel Ai Learning, 5 Technical University of Munich
Preprint
Code Paper
GIF 1
DE-LU region
GIF 2
FR region
GIF 3
NL region

Example forecasting on testing samples

Abstract

Electricity price forecasting in Europe presents unique challenges due to the continent's increasingly integrated and physically interconnected power market. 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 crossregion dependencies in interconnected power systems, motivating a domain-specific foundation model. In this paper, we address this gap by first introducing a comprehensive and up-to-date dataset across 24 European countries (38 regions), spanning from 2022-01-01 to 2026-01-01. Building on this groundwork, we propose PRICEFM, a probabilistic foundation model pretrained on this large dataset. Specifically, PRICEFM maps each region's price and exogenous features into a comparable latent embedding via a shared Mixture-of-Experts (MoE) projection layer, then injects prior graph knowledge by constructing a sparse graph mask derived from transmission topology. Across a large-scale European benchmark, PriceFM achieves strong performance and demonstrates superior generalization under both zero-shot and full-shot evaluation compared with multiple competitive baselines.

Teaser Image

Structure of PriceFM. The input features are passed into a MoE projection layer to produce the regional representations. The regional representations are stacked to form the shared spatial representation S, which is multiplied with the sparse graph mask to produce the spatial representation, which is then fed into hierarchical quantile heads to produce probabilistic forecasts.

Energy data

European-level energy data in 2025, averaged across regions. a Electricity price. Price spikes sharply during the morning and evening peak, dip around midday. b Forecasted load. Load exhibits a double-peak each day. c Forecasted solar power generation. Solar is zero overnight, rises in a smooth bell curve to a strong midday maximum, then falls back to zero by dusk. d Forecasted wind power generation (onshore and offshore). Wind lacks a daily pattern, fluctuates with high-frequency spikes.

Model comparison results

Results of model comparison.

BibTeX

If you use our code or find our paper useful, please cite:

@misc{YuPriceFM,
      title={PriceFM: Foundation Model for Probabilistic Electricity Price Forecasting}, 
      author={Runyao Yu and Chenhui Gu and Jochen Stiasny and Qingsong Wen and Wasim Sarwar Dilov and Lianlian Qi and Jochen L. Cremer},
      year={2026},
      eprint={2508.04875},
      archivePrefix={arXiv},
      primaryClass={cs.CE},
      url={https://arxiv.org/abs/2508.04875}, 
}