The Australian Energy Market Commission (AEMC) published its long-delayed report on risks of algorithmic collusion in battery trading.
The report warns that AI tools and increased market data create ideal conditions for tacit collusion, where algorithms align on higher prices without explicit communication.
Detecting such collusion would be prohibitively expensive, while retraining algorithms or restricting auto-bidding is considered impractical.
Big batteries, once expected to curb high prices, are increasingly controlled by dominant market players and already driving price-setting in the NEM.
AEMO data showed battery discharges set prices at $478/MWh in June 2024, nearly double the previous year and triple gas prices.
The AEMC says home batteries could disrupt collusive behaviour by injecting supply when prices spike, reducing AI-driven profits.
Hedging contracts provide some short-term protection, but sustained manipulation would drive these costs up.
Current NEM rules around transparency and bidding offer partial safeguards, but monitoring black-box AI systems remains a major challenge.
The AEMC recommends giving the ACCC greater information-gathering powers and forming a joint working group with regulators to tackle the issue.
Market barriers such as transmission congestion, long wait times for new projects, and rising generation costs further heighten the risk of AI-enabled collusion.