Energy Storage Summit USA: Here’s What I Came Back With

This week I attended the Energy Storage Summit in Dallas and I went in with a simple question: why is BESS economics so hard to get right, and where does better data fit in? Battery storage should be a straightforward value proposition, but the reality is messier than the model — and most operators I’d talked to felt like they were leaving money on the table without knowing exactly why. These are my takeaways. 

With the event being hosted in Dallas, it had the feel of a regional event weighted heavily toward ERCOT — and the mood on the floor reflected it. The BESS business in TX has had two tough years: ancillary service saturation, back-to-back mild summers, and a drought of the scarcity events that made 2022–2023 so lucrative. 

Average ERCOT battery revenues fell from $140–200/kW-year at peak to as low as $30 in 2025. But the attendees were optimistic — accelerating load growth from data centers, an inevitable return of extreme weather, and the fundamental irreplaceability of storage on a renewables-heavy grid all point in one direction. The consensus: lean years are the right time to close the gap between what your assets are theoretically capable of and what they deliver. They shared some insights on that front that a Digital Control Platform like Ardexa’s directly address: 

The Revenue Gap Is a Data Problem 

Every panel kept returning to the same finding: the gap between modeled and actual BESS revenue isn’t just bad market forecasting — it’s unreliable BESS data. BMS errors at the rack level aggregate up to inaccurate state-of-charge signals, and optimizers forced to apply wide safety buffers because they can’t trust what the asset is reporting. Fix the data layer, and you start closing the gap. 

The Safety Buffer Tax Is Costing Operators a Fortune

SOC signals flowing to BESS optimizers are often significantly inaccurate. Optimizers can’t trust what they’re seeing, so they do the rational thing: add large safety buffers, routinely limiting dispatch to just 20–25% of actual remaining capacity to avoid unexpected derating events. The math is brutal — a site with 19.5 MWh of accessible energy may only ever see 7 MWh monetized. That isn’t a market problem. It’s a data problem, and it’s solvable today. 

The Perfect Foresight Gap Has Three Sources

Panelists identified three compounding causes of revenue underperformance: inaccurate price forecasts, suboptimal dispatch, and technical underperformance of the asset itself. Forecasters and Bidders can only address the first two if the third is solved. No bidding algorithm recovers revenue that was never physically available because state-of-charge was misreported or a rack was quietly underperforming. Accurate high-frequency asset data is what makes the optimizer’s job possible.