MELISANDRE : Market Equilibria Likelihood and Safety Assessment in AI-Driven energy systems

Recipient
Georgios Tsaousoglou
Technical University of Denmark
Project number:
00080181
Grant amount
8.199.494 DKK
Year
2025

Project description

Power systems are the economy’s backbone. It is crucial to operate them efficiently and reliably. This is pursued through electricity markets, meant to ensure that electricity demand is met at minimum cost. However, recent crises make it clear that we are entering unknown territory, while the most concerning change is just beginning to unfold:

Market participants are increasingly using AI agents for their bidding strategies. 

While AI agents are not trying to harm the system, each does try to maximize its own profits, possibly in ways we didn’t anticipate. This disrupts the previous market patterns, as we don’t yet understand whether the interaction of such AI-powered profit-driven strategies can potentially lead to unforeseen and detrimental system outcomes. MELISANDRE assesses the range and likelihood of new market equilibria and designs interventions that steer equilibria away from regions that pose risks to system efficiency and safety.