Battery energy storage has scaled up at remarkable speed. McKinsey expects global BESS capacity to rise from around 200 GWh at the end of 2025 to between 500 and 700 GWh by 2030, nearly tripling in five years. New sites are coming online faster than the specialist teams running them can keep up with their complexity. There is no shortage of information. If anything, every site produces too much of it: battery management system (BMS) outputs, energy management system (EMS) logs and operator dashboards, plus a constant stream of alarms, flags and notifications that no team can process in real time.
A bankable battery needs a digital exit strategy
It is somebody’s job to sift through hundreds of alarms to find the handful that matter, and doing that well takes deep knowledge of each site. That may work for one site, but it does not work for two, or three, or an entire portfolio. In practice, the most severe issues get resolved quickly, while seemingly minor ones are left alone until they escalate, eating into revenue in the meantime. And almost none of these signals are predictive. An alarm may tell you a value is out of range right now; it says nothing about what is coming. The systems on site keep you informed, but they do not keep you ahead.
Between alarms, operators manage that uncertainty the only way they can, by building in safety margins. Margin on available power, on maintenance intervals, on the revenue they commit to. Each buffer is an educated guess, sized for a risk nobody can reliably measure. Over time, these margins turn into money left on the table, built into how the asset is run.
What proactive management looks like
For operations and maintenance, getting ahead of problems takes two things. First, operators need to get alert fatigue under control and work out what actually needs attention. Second, they need to look ahead. The first is easy to measure. In my experience, a monitoring layer that draws on all the data a site produces and evaluates signals in context typically cuts the number of alarms a team has to review by a factor of ten to twenty. Instead of scanning hundreds of alarms, you focus on a few, ranked by how much fixing them would improve the asset’s performance.
Giannis Giannakopoulos: “Portfolio and aggregator management is essential”
Looking ahead is a question of timescales. Long-range forecasts, such as how a cell ages or where its state of health is expected to land at end of life, matter enormously for financing, augmentation planning and warranty cases. Daily operations, however, run on shorter horizons: which string will limit power after a certain throughput, which imbalance is starting to drift, what needs attention in the coming days and weeks.
Once operators can act early, the benefits reach beyond the control room. Each issue caught in time removes a risk they would otherwise carry without knowing its size. Lenders and insurers can price an asset more accurately when its failure modes are understood and prevented. Some already offer better terms for assets under continuous monitoring. Asset managers report numbers they can trust, and the operations team ends up with something close to a plan: a ranked view of what needs attention, when and why.
Where signals you can trust come from
Much of the monitoring on site comes from the manufacturers, and it reflects what each manufacturer knows: its own hardware, its own lab tests, its own assumptions. It sees the operator’s data only selectively, if at all, and it never sees how other vendors’ equipment behaves in the field.
An independent platform works from a different foundation. The data belongs to the owner and operator, and it spans manufacturers, use cases, operating modes and markets. At ACCURE, more than 24 GWh of storage across the US, Europe and Australia feeds a single analytics platform. Our laboratory, in effect, is the field. That breadth makes insights transferable. A failure mode that first appears on assets in the Texas heat can be recognised early on Australian sites facing similar conditions, before it causes damage. And because the platform is software, with no hardware cycle to wait for, the road from a new insight to a live warning across an entire portfolio is short.
Europe has built the batteries, and now has to run them well
The ACCURE Safety Score is a good example of what this looks like in practice. The score is a single, continuously updated estimate of the risk that a rack fails or, worse, turns into a safety incident. Instead of forwarding every small deviation or anomaly, the platform tracks hundreds of signals over days, watches how they develop and which ones occur together, and condenses hundreds of thousands of data points into one value. Only when that value rises, and the projection shows the risk will keep building, does the operator get an alert. It comes with a risk estimate, enough lead time to plan the intervention and clarity on who bears the cost.
Get ahead
A warning only helps if it arrives before the damage is done, early enough to leave operators in control of when to act and what to prioritise. Nor does early warning stop at the battery itself. It extends to the heating, ventilation and air-conditioning (HVAC) system, the power conversion system and the rest of the balance of plant. The data makes it possible, and the models get sharper with every day of operation.
A site or two can survive on a reactive cycle, but a portfolio of gigawatts spread across continents cannot. At scale, the performance gap between operators who run proactively and those still working through a backlog of alerts starts showing up on the balance sheet. (Valentin Lorscheid/hcn)
