Post: Your revenue forecast assumed the battery would perform. Does your platform make sure it does?

Your revenue forecast assumed the battery would perform. Does your platform make sure it does?

Guest post by GridBeyond Business Development Director (AU), Mark Netto

FCAS revenues have compressed and everyone in the market knows it. But exiting FCAS entirely is a mistake; and it’s not the biggest operational risk facing sub-5MW battery projects right now. The bigger issue is the gap between what a project modelled at financial close and what it actually earns.

Whilst any gap is likely to have some forecast error associated with it, it is far more likely to be driven by:

• cycling decisions made without knowing the true cost
• rack imbalance going undetected
• SoC data that isn’t accurate enough to trust

The platform you choose determines how much of that gap you close.

There’s a version of the FCAS conversation that is now mostly background noise. Contingency FCAS revenues have compressed. The market has become more competitive, asset counts have grown, and the returns available have reduced accordingly. But FCAS has always been the cream on top; additional revenue layered over the energy arbitrage case, not the foundation of it. Most developers building sub-5MW projects today already price it that way.

What is less well understood is why FCAS participation still matters at the margin and why the operational decisions sitting underneath it have become the real driver of long-term project performance.

You need to be in it to win it

A battery that exits FCAS markets entirely to avoid competitive pressure is leaving money on the table during the intervals between energy volatility events. Those periods are not short; they make up the majority of operating hours across a year. FCAS keeps the asset earning through them.

There is a more practical point too. Assets operating without active market participation are also operating without active oversight and that is where the real risk lies.

We have seen cases where a connection between the site and the platform has dropped out silently. No alert. No visible failure. The asset appears operational from the outside, but the optimiser has lost telemetry and is no longer making informed decisions. Without continuous someone watching the platform around the clock and tracking whether data is actually flowing, that kind of failure can go undetected for hours. During a high-price event, that is a material revenue loss. During an FCAS enablement period, it is a compliance exposure.

The value of keeping the asset actively participating in the market is partly about the revenue. It is also about maintaining the continuous monitoring loop that catches these failures before they become costly.

Every cycle has a cost. Most platforms don’t know what it is.

This is where the conversation gets more interesting; and where the gap between high-performing and average battery portfolios is increasingly being set.

Every dispatch decision has a true cost. That cost is not fixed. It is a function of the battery’s state of health (SoH), its depth of discharge, its chemistry, and its age. For an LFP battery, the relationship between depth of discharge and cycle life is non-linear. At shallow depths of discharge the battery can sustain a very high number of lifetime cycles. As depth of discharge increases, the number of full cycle equivalents the battery can perform over its lifetime decreases and the marginal cost of each cycle increases accordingly. A dispatch decision made at 80% depth of discharge is materially more expensive, in lifetime value terms, than the same decision made at 40%.

A platform that doesn’t integrate real-time SoH and depth of discharge into a live marginal cost calculation is making every dispatch call without knowing what it actually costs. It will cycle when the revenue doesn’t justify the degradation. It will hold back when it does. Over time, that adds up.

The flip side is equally important. During a genuine scarcity event (prices above $5,000/MWh) the right decision may be to deliberately accept additional degradation. Extra aging can be justified when the revenue is exceptional. A platform that knows its true marginal cost can make that trade explicitly and confidently. One that doesn’t will either always avoid it or never avoid it, neither of which is optimal.

The rack problem nobody talks about

Battery systems are not monolithic. They are made up of racks—individual battery modules that collectively determine total system output.

Rack imbalance is one of the most common and least discussed sources of revenue leakage in operating battery projects. It develops gradually, through uneven usage patterns, calibration drift, and sensor errors. And it has a disproportionate impact on system performance.

A single imbalanced rack will reduce the output of the otherwise available and healthy racks it is connected to. For example, a rack with an 11% imbalance will pull the performance of the healthy racks down to its level. Meaning the inverter, the connection point, and every other part of the system may be fully available and healthy, but total dispatchable output is constrained by the weakest point. In FCAS terms, this creates delivery risk. In energy terms, it means the asset is consistently underperforming against its contracted or modelled capacity.

The problem is that rack imbalance is largely invisible without continuous rack-level monitoring. It doesn’t show up in aggregate SoC data. It often doesn’t appear in O&M reports until the imbalance is significant enough to trigger isolation; at which point revenue has been lost and degradation has been accelerated for some time.

Proactive monitoring of SoC and voltage across individual racks in real time is what catches this early. Recalibration or targeted servicing of underperforming racks maintains full usable capacity. The difference between detecting a problem at 3% imbalance and discovering it at 11% is meaningful in revenue terms across a project lifetime.

Garbage in, garbage out

Every layer of the optimisation stack described above depends on accurate state of charge data.

Marginal cost calculations are built on SoC. FCAS availability offers are built on SoC. Rack imbalance detection is built on SoC. If the underlying SoC measurement is drifting due to sensor error, calibration issues, or inadequate measurement methodology, every decision downstream of it is less reliable than it appears.

An LFP battery’s flat voltage curve during mid-range charge states makes accurate SoC estimation genuinely difficult. It requires more than a simple voltage reading, it needs current integration, temperature correction, and regular calibration anchoring. A platform that treats SoC as a given rather than something to be continuously verified will accumulate error over time, and that error will show up in dispatch decisions, FCAS delivery, and degradation outcomes.

Improved SoC measurement is not a minor technical detail. It is the foundation the rest of the optimisation logic is built on.

The gap between modelled and realised

The battery projects that underperform against their financial models rarely do so in isolation from forecast error, but whilst that plays a role, the gap between modelled and realised revenue is far more likely to be driven by the same operational sources:

• cycling decisions made without accurate marginal cost data
• rack imbalance that went undetected for too long
• SoC drift that quietly degraded the quality of every optimisation decision

None of these are exotic failure modes. They are the normal operational reality of battery assets over a multi-year project life.

The question is whether the platform managing the asset is actively managing against them or whether it is simply dispatching and reporting.

As the market matures and institutional capital becomes more sophisticated about what it is actually funding, these distinctions are going to become harder to ignore.

If you’re currently evaluating platforms, building a revenue model, or trying to understand why assets underperform against financial close assumptions, happy to compare notes.

Series note

This is the fourth in a series on what separates high-performing sub-5MW battery portfolios.

Previous articles explored:

• the core capabilities required to operate sub-5MW portfolios
• forecasting performance as the foundation of optimisation
• hybrid optimisation and where value is lost

In the next article, I’ll look at how consolidating optimisation and market participation under a single provider simplifies operations and why that’s becoming an increasingly important consideration for sub-5MW portfolio developers.