Battery Energy Storage Trading & ETRM

Executive summary

Battery energy storage has moved from the edge of power markets to their centre. Grid-scale batteries now provide arbitrage, ancillary services, congestion relief, and firming for renewables, and they are being built at a pace that is reshaping how electricity markets clear. Yet very few ETRM platforms model batteries properly, because a battery is not a static position, it is a trading instrument that changes state every interval and demands continuous optimisation.

A battery is genuinely different from a generator or a contract. Its value depends on its state of charge, its power and energy ratings, its round-trip efficiency, its cycle limits, and the degradation each cycle imposes, all against an intraday price shape that moves constantly. Extracting that value is an optimisation problem across energy, ancillary, and capacity revenue, at interval granularity, that a block-oriented system simply cannot express.

This guide covers the complete picture: why batteries transform markets, the trading lifecycle, the battery asset model, energy arbitrage, ancillary services, virtual power plants, dispatch optimisation, risk and valuation, and real-time operations. It builds on the power trading lifecycle and connects to the AI-and-congestion analysis, scenario analysis, and the power commodity page.

Why batteries transform electricity markets

Batteries change the shape of power markets because they can move energy across time, which non-storable electricity otherwise cannot. A battery charges when power is abundant and cheap, often when solar floods the midday market, and discharges when it is scarce and expensive, typically the evening peak, smoothing the price shape and capturing the spread in between.

That single capability makes batteries central to several market functions at once: energy arbitrage across the day, firming and integrating variable renewables, providing fast ancillary services such as frequency regulation, relieving congestion when sited behind a constraint, and contributing firm capacity. As renewables grow and AI-driven demand steepens the intraday shape, the value of this flexibility rises, which is why batteries are being deployed so rapidly and why trading them well is increasingly a competitive necessity for power desks.

The battery trading lifecycle

Trading a battery follows a lifecycle, but one that repeats continuously rather than running once per deal, because the optimal charge and discharge plan is re-solved as prices and forecasts change.

The distinguishing feature is that optimisation is continuous and multi-market. A battery’s value is not set at trade time; it is realised interval by interval as the desk decides, repeatedly, whether each megawatt and each unit of stored energy is worth more charging, discharging, or held for an ancillary or capacity commitment. This is why a battery demands a platform that re-optimises against live data rather than one that records a static position.

The battery asset model

Everything starts with modelling the asset faithfully. A battery is defined by a set of physical parameters that bound what it can do and determine its value.

These parameters are not background detail; they are the constraints the optimisation runs against. State of charge determines what the battery can do next; efficiency losses mean the spread has to exceed a threshold to be worth cycling; cycle limits and degradation mean a battery dispatched without accounting for wear can earn revenue today while destroying asset value tomorrow. A platform that models these natively, at interval granularity, is the prerequisite for trading a battery properly.

Energy arbitrage

The foundational battery strategy is energy arbitrage: charge when power is cheap, discharge when it is dear, and capture the spread. The value depends on the intraday price shape, the battery’s duration and efficiency, and how many cycles the asset can afford.

Arbitrage is more subtle than buy-low-sell-high because efficiency losses and degradation impose a real cost on every cycle. A spread has to clear that cost to be worth capturing, and a battery that cycles indiscriminately to chase small spreads can lose money once wear is counted. Locational arbitrage adds another dimension: a battery sited behind a transmission constraint can capture locational spreads, driven by congestion, that a battery elsewhere cannot. Modelling arbitrage properly therefore requires the intraday shape, the asset constraints, and the nodal, congestion-aware view together, on one model.

Ancillary services and co-optimisation

Energy arbitrage is only part of a battery’s value. Batteries are exceptionally good at fast grid services, and often earn more from ancillary markets than from energy alone. The relevant products include frequency regulation and response, operating reserves, and, in some markets, firm capacity commitments.

The key insight is co-optimisation. In any interval, a megawatt of battery capacity can be sold into the energy market, offered as a reserve or frequency-response product, or held against a capacity commitment, and these choices compete for the same physical asset. Optimising them together, rather than in sequence, is what maximises total value, and it only works if all the relevant markets sit on one model. A battery managed for energy arbitrage alone, ignoring the ancillary stack, typically leaves substantial revenue unclaimed.

Virtual power plants and DER aggregation

Batteries are increasingly traded not as single assets but as fleets. A virtual power plant (VPP) aggregates many distributed batteries, and often other distributed energy resources such as demand response and flexible load, into a single dispatchable resource that trades as one entity while respecting the physics of each unit.

Aggregation raises the modelling bar. Dispatch has to honour each unit’s state of charge, power rating, and location while presenting a coherent aggregate position to the market. The locational dimension matters too, since units behind different constraints have different value. A platform that can hold a fleet of distributed assets on one governed model, optimising the aggregate while respecting each unit, is what makes VPP and DER aggregation tractable rather than a spreadsheet nightmare.

Dispatch optimisation

At the heart of battery trading is dispatch optimisation: given forecasts, prices, market rules, and the asset’s constraints, decide the charge, discharge, and hold schedule that maximises value. This is a genuine optimisation problem, not a rule of thumb, and it has to be re-solved as conditions change.

A good optimisation respects everything at once: the intraday price shape and its uncertainty, the battery’s state of charge and power and energy limits, round-trip efficiency, cycle limits and degradation cost, and the competing value of energy, ancillary, and capacity markets. The output is a dispatch plan the desk can bid and execute, updated as forecasts move. Because the inputs are all governed data on one model, the optimisation is consistent with the position, the risk, and the settlement that follow, which is what makes its recommendations trustworthy rather than a black box.

Risk and valuation

Valuing a battery is harder than valuing a forward, because its worth is the expected profit from optimally operating it against an uncertain price shape, net of efficiency losses and degradation, subject to its constraints. In effect a battery is a bundle of options on the intraday and inter-market spreads, and its value moves with the forward curve and volatility.

That valuation feeds the same risk stack the rest of the desk uses. A battery position carries price and shape risk through the intraday spread, volume risk if it cannot deliver as planned, and operational risk from the asset itself, and this exposure belongs in portfolio VaR and scenario analysis alongside everything else. Because battery value depends on the shape of the forward curve, it contributes to the portfolio’s sensitivity to curve and volatility moves, and managing it on one governed model is what keeps the firm’s risk picture honest.

Real-time operations

Batteries are operated in real time, and the trading platform has to keep pace. Dispatch instructions execute against live telemetry, state of charge updates continuously, and the optimisation re-solves as prices and forecasts move, all within the interval timescales the market runs on.

This demands an event-driven architecture: SCADA and IoT telemetry feed the current state of charge and availability; market data and forecasts drive re-optimisation; and the position, risk, and P&L update as the battery cycles. A trader or operator sees the live state, the current plan, and the exposure at any moment, rather than an end-of-day reconstruction. This real-time loop, forecast, optimise, dispatch, settle, running continuously on one governed model, is what battery trading actually requires and what a modern platform provides.

Best practices

Trading batteries well rests on a few principles. Model the asset natively, state of charge, efficiency, cycle limits, and degradation, at interval granularity, so the optimisation runs against real constraints. Co-optimise across energy, ancillary, and capacity markets rather than chasing energy arbitrage alone. Account for degradation as a real cost so the battery is not cycled to destruction for thin spreads. Value the battery as the option bundle it is, and carry that exposure into portfolio risk. And run the whole loop, forecast, optimise, dispatch, settle, in real time on one governed model.

The through-line is that a battery is a continuous, multi-market optimisation problem, not a static position. The platforms that trade batteries well are the ones designed for granularity, state, optionality, and real-time optimisation from the start, rather than retrofitting shape onto a coarser, block-oriented design.

Operational KPIs

A battery trading operation can be measured across value capture and asset health.

Revenue versus optimal measures how much of the achievable value the desk captures; cycles versus warranty and degradation tracking measure whether value is being taken without destroying the asset; state-of-charge accuracy and dispatch execution measure real-time control. Together they describe a battery run as the value-generating, finite-life asset it is.

Why the Gravitas battery module is different

Gravitas models batteries natively, at interval granularity, on one governed model.

Because the asset model, the optimisation, the position, the risk, and the settlement all sit on one model, a battery is traded as the continuous, multi-market optimisation it is, with its value captured and its asset life protected. And it is delivered at economics that suit desks the incumbents priced out. See the power commodity page, who Gravitas is for, or request a demo.