Battery Bidding Algorithms Could Reshape Storage Revenue

A new preprint from arXiv models how battery storage operators should bid into wholesale electricity markets when future prices are uncertain. The study, titled "Battery Bidding under Price Uncertainty in Wholesale Electricity Markets," finds that the standard approach – bidding based on expected price spreads – misses revenue when price volatility is high. A strategy that accounts for the full distribution of possible prices can capture significantly more revenue, the paper concludes.

The research arrives as grid-scale battery capacity expands. The U.S. Energy Information Administration reported developers added more than 10 gigawatts of storage in 2024 alone. For operators of these assets, the bidding algorithm is a direct lever on profitability. A battery that charges when power is cheap and discharges during peak hours can see returns swing by 20 percent or more depending on how well bids are calibrated to real-time price distributions.

The paper's authors, affiliated with a major university energy lab, tested their model on historical price data from California's CAISO market. They showed that a bid function accounting for both expected price and variance outperformed a simple spread-based strategy by roughly 15 percent in simulated revenue. That gap widens in markets with high renewable penetration, where intraday price swings are larger.

For listed companies in the battery storage space – including Fluence Energy and Tesla Energy – the finding has direct commercial relevance. Both firms sell energy management software that handles bidding into wholesale markets. Fluence's AI-powered trading platform, for instance, is marketed as a way to optimize revenue from storage assets. If academic models continue to refine bidding under uncertainty, the competitive advantage may shift toward operators who can integrate these newer algorithms into their trading engines.

The immediate catalyst for energy storage stocks is not the paper itself. It is the ongoing regulatory push to align battery market participation with FERC Order 841, which requires grid operators to allow storage to compete in wholesale markets. As more independent storage owners enter these markets, the demand for sophisticated bidding software rises. The arXiv research suggests the gap between a naive bid and an optimized bid is large enough to justify buying the better software.

One risk: the paper is theoretical. Real-world implementation depends on accurate price forecasts and low-latency computation. A battery operator using a flawed volatility estimate could underperform the spread bid. Still, the direction of the research is clear – bidding algorithms that price uncertainty will become a standard part of storage operations. The software vendors that embed them early may lock in customers before competitors catch up.

The paper has not yet been peer-reviewed for publication. It is posted as a preprint on arXiv, which means the conclusions could change after formal review. For traders watching the energy storage sector, the next concrete event is the U.S. Department of Energy's workshop on grid-scale battery market design, scheduled for June.