The Moment Energy Arbitrage Stopped Being Enough
In 2017, a 100 MW lithium-ion battery began operating in South Australia. It was, at the time, the largest grid-connected battery in the world. What received less attention was how it made money. The original business case relied on battery storage revenue stacking — frequency control and ancillary services, not buying low and selling high. Energy arbitrage alone could not justify the capital expenditure. That reality has shaped the economics of grid-scale storage ever since.
Most people outside the electricity industry still imagine batteries as giant versions of the one in a mobile phone: charge when electricity is cheap, discharge when it is expensive. That is technically possible. It is also, in almost every market today, financially insufficient. A battery that only performs energy arbitrage will struggle to recover its investment within a reasonable timeframe. The business case for grid-scale storage relies on something the industry calls revenue stacking — combining multiple income streams from different markets and services, often within the same hour.
Revenue stacking is a structural feature of how batteries create value in electricity systems. Understanding which revenue streams exist, how they fit together, and what constraints limit their combination is essential for anyone trying to understand why batteries are being deployed where they are — and why some markets have seen far more activity than others.
Energy Arbitrage: The Foundation, Not the Building
Energy arbitrage is the simplest revenue stream to explain. A battery charges during periods of low wholesale electricity prices and discharges during periods of high prices, capturing the spread. In markets with significant solar penetration, this often means charging during midday when solar generation suppresses prices and discharging during the evening peak when demand rises and solar output falls. The pattern is predictable enough that it has acquired its own shorthand: the duck curve trade.
The economics are less straightforward than they appear. A lithium-ion battery system operating in wholesale markets typically achieves round-trip efficiency of 85 to 90 percent. For every 100 MWh of electricity that enters the battery, approximately 85 to 90 MWh can be sold back. The price spread must be wide enough to cover those efficiency losses, plus the degradation of the battery cells, plus the cost of capital. In practice, spreads of 50 to 70 percent between charge and discharge prices are often necessary simply to break even on energy arbitrage alone.
Markets with deep solar penetration have produced the most attractive arbitrage opportunities. California’s CAISO market regularly sees midday prices fall to near zero — or occasionally negative — while evening prices rise sharply. Australia’s National Electricity Market exhibits similar patterns, with the added complication of more volatile weather-driven price spikes. But even in these favourable environments, arbitrage revenues alone have typically fallen short of what developers require to finance projects. BloombergNEF analysis has consistently shown that standalone arbitrage revenues in most major markets cover only a portion of a battery’s levelized cost of storage. The remaining revenue must come from elsewhere.
Frequency Response: Speed as a Revenue Product
This is where the technical characteristics of batteries become commercially decisive. Unlike thermal generators, which take minutes to adjust output, a battery can respond to a frequency deviation in milliseconds. That speed has economic value because grid frequency must be maintained within narrow tolerances — typically 50 Hz or 60 Hz, depending on the region — and deviations must be corrected almost instantly to prevent equipment damage or cascading failures.
The United Kingdom’s National Grid was one of the first system operators to create a market specifically for fast frequency response. The Enhanced Frequency Response service, introduced in 2016, paid batteries a fixed availability fee for standing ready to inject or absorb power within one second of a frequency deviation. The contracts were awarded through competitive tenders and provided a predictable revenue stream that fundamentally changed the investment case for battery storage in Britain. By 2020, batteries had become the dominant provider of frequency response services in the UK market, displacing gas-fired plants that were slower and more expensive to operate in that role.
Australia’s Frequency Control Ancillary Services markets tell a similar story. The Hornsdale Power Reserve — the South Australian battery mentioned earlier — earned a significant share of its early revenues from frequency control. In 2018, its first full year of operation, frequency control services accounted for an estimated 50 to 60 percent of total revenue, according to analysis by the Australian Energy Market Operator and independent research firms. Energy arbitrage contributed less than 20 percent. The pattern has shifted somewhat as more batteries have entered the market and competition for frequency services has intensified, but the principle remains: speed is a monetizable asset.
What makes frequency response particularly attractive for battery operators is that it does not require deep cycling of the battery. A frequency response event typically requires only a small injection or withdrawal of power — often one to ten percent of the battery’s rated capacity — for a duration measured in seconds or minutes. This preserves the battery’s state of charge for other uses and minimizes degradation. A battery providing frequency response can simultaneously participate in energy markets, provided the system is sophisticated enough to manage the competing demands.
Capacity Markets and Resource Adequacy Contracts
The third major revenue stream comes from being paid simply for existing. Capacity markets and resource adequacy contracts compensate generators — and increasingly, storage assets — for committing to be available when the grid needs them most, typically during periods of peak demand or system stress. The payment is for availability, not for energy delivered. For a battery, this means receiving a steady revenue stream for guaranteeing that a certain amount of power can be dispatched when called upon, regardless of whether it actually is.
Capacity market design varies significantly across jurisdictions, and those differences have a material impact on battery economics. The United Kingdom’s Capacity Market runs auctions four years ahead of delivery, providing revenue visibility that supports project financing. Battery storage projects have been successful in these auctions, typically securing one-year agreements at clearing prices that have ranged from approximately £15 to £30 per kilowatt per year in recent auctions. A 100 MW battery securing a contract at £20/kW/year would receive £2 million annually in capacity payments before ever discharging a single megawatt-hour of energy.
In the United States, PJM Interconnection operates one of the world’s largest capacity markets, and battery storage participation has grown substantially. But the market design presents challenges for storage. PJM’s capacity product traditionally required assets to demonstrate the ability to run for extended durations — often ten hours or more — which exceeds the typical duration of a lithium-ion battery system. Regulators and market operators are adjusting these requirements as storage penetration increases, but design details matter enormously. A battery capable of four hours of continuous discharge may receive only a fraction of the capacity credit that a gas plant of equivalent nameplate capacity would receive, because the market values duration differently.
California has taken a different approach through its Resource Adequacy program, which requires load-serving entities to contract for sufficient capacity to meet peak demand plus a reserve margin. Battery storage has become a major beneficiary of this framework. Four-hour duration batteries — which align closely with the shape of California’s evening peak — have contracted extensively with utilities and community choice aggregators. The revenue from these resource adequacy contracts often provides the anchor tenant for project financing, with arbitrage and ancillary services contributing additional upside.
The Operational Reality: Stacking Without Compromising
The cumulative case sounds straightforward: combine arbitrage, frequency response, and capacity payments and the numbers work. The operational reality is considerably more complex. A battery cannot simultaneously be fully charged for arbitrage, partially discharged for frequency response, and held in reserve for a capacity obligation without sophisticated management. Each revenue stream places different demands on the state of charge, and the system must optimize across them in real time.
The degradation trade-off is particularly significant. Lithium-ion batteries degrade with use, and the rate of degradation depends on how they are used. Deep cycling — regularly charging and discharging across the full state-of-charge range — accelerates degradation more than shallow cycling. A battery optimized aggressively for arbitrage revenue might cycle deeply twice per day, maximizing spread capture but shortening the asset’s useful life. A battery operated conservatively for frequency response might degrade far more slowly but leave significant arbitrage revenue uncaptured. The operator must decide, continuously, where to position the battery along this spectrum.
This optimization problem has spawned a sub-industry of battery management software and algorithmic trading platforms. Companies like Habitat Energy, Flexitricity, and Tesla’s Autobidder have developed systems that forecast prices across multiple markets, predict frequency response requirements, and dispatch the battery accordingly. The quality of this software can materially affect project returns. Two batteries with identical hardware, located in the same market, can produce substantially different revenues depending on the sophistication of the optimization algorithm controlling them. This is not a minor detail: revenue differences of 10 to 20 percent attributable to trading strategy have been observed in markets where independent operators compete.
Regulatory constraints add another layer of complexity. Not all markets allow a single asset to participate in multiple revenue streams simultaneously. Some system operators require batteries providing frequency response to reserve a certain amount of capacity for that purpose, reducing their ability to participate in energy markets. Others restrict the combination of capacity market obligations with merchant trading. Market design choices made years ago — often before battery storage was contemplated at scale — continue to shape the revenue stacking possibilities available today. The most successful battery markets have been those where regulators and system operators have explicitly designed market rules to accommodate storage’s unique characteristics.
Market Data: Where the Revenue Actually Comes From
The composition of battery revenues varies significantly by market, and the patterns have shifted as more storage has been deployed. In the United Kingdom, analysis by Modo Energy — a specialist battery storage data platform — shows that Dynamic Containment, a fast frequency response product, has accounted for a substantial share of battery revenues since its introduction in 2020. As more batteries have entered this market, revenues from Dynamic Containment have declined, pushing operators to diversify into wholesale trading and the Balancing Mechanism. This pattern — early entrants capturing high ancillary services revenues, followed by saturation and diversification — has become characteristic of maturing battery markets.
In ERCOT, the Texas market, battery revenues have been dominated by energy arbitrage and ancillary services, with particularly strong performance during periods of extreme weather. During Winter Storm Uri in February 2021, wholesale electricity prices reached the market cap of $9,000 per megawatt-hour. Batteries capable of discharging during those hours captured extraordinary revenues — and the event reshaped market expectations about battery economics in deregulated markets. Subsequent regulatory reforms, including changes to the operating reserve demand curve and ancillary services procurement, have further influenced the revenue stacking calculus.
Australia’s National Electricity Market provides perhaps the clearest illustration of revenue stacking in practice. The Australian Energy Regulator’s wholesale market data shows battery revenues typically split across energy arbitrage (capturing spreads between solar troughs and evening peaks), frequency control ancillary services, and — increasingly — network support contracts. The proportions shift both seasonally and structurally as the battery fleet grows. In 2022, frequency control revenues were estimated to account for roughly 40 to 50 percent of total battery revenue in the NEM, with energy arbitrage contributing 30 to 40 percent and the remainder from other services. By 2024, those proportions had reportedly shifted toward a more even split as ancillary services markets became more competitive.
Revenue stacking is an operational strategy that must adapt as market conditions change. The battery that was profitable on frequency response in 2020 may find those revenues declining by 2024 and must shift toward energy trading. The operator that can adapt its strategy fastest captures the remaining value. This is a fundamentally different business model from traditional generation, where the revenue model — produce energy, sell energy — changes only at the margins over an asset’s life.
What Happens Next
The revenue stacking model faces several structural challenges as battery deployment accelerates. The most significant is market saturation. Frequency response and ancillary services are relatively small markets — typically measured in hundreds of megawatts, not gigawatts. As more batteries enter these markets, revenues per megawatt decline. This has already occurred in the UK frequency response market and is beginning to occur in other jurisdictions. Batteries that entered early captured high ancillary services revenues; late entrants face a more challenging environment and must rely more heavily on energy arbitrage and capacity payments.
Longer-duration storage technologies — flow batteries, compressed air, and eventually hydrogen — may shift the revenue stacking calculus further. A battery capable of eight or twelve hours of discharge can access different revenue opportunities than a two-hour or four-hour system, including longer-duration capacity products and multi-day arbitrage strategies. However, these technologies currently cost more per kilowatt-hour of storage capacity than lithium-ion, and the revenue stacking math for longer-duration assets remains unproven in most markets.
For now, the day-to-day economics of battery storage remain rooted in the principle that no single revenue stream is sufficient. A typical project stacks them: energy arbitrage covers the base return, frequency response and ancillary services add the premium, and capacity payments or resource adequacy contracts provide the floor that supports the financing. The art — and increasingly, the algorithmic science — lies in combining them optimally. The operators who master that combination will determine where batteries get built, how they get financed, and ultimately, how they reshape the economics of electricity systems around the world.
References
- IEA — World Energy Outlook and Battery Storage reports providing context on global battery deployment trends and cost trajectories
- BloombergNEF — Battery storage cost surveys and revenue analysis across major markets, particularly UK, Australia, and California
- Australian Energy Market Operator (AEMO) — Wholesale market data and Hornsdale Power Reserve operational reports illustrating frequency control revenue shares
- Modo Energy — UK battery storage revenue indices and Dynamic Containment market analysis tracking revenue composition shifts
- Australian Energy Regulator — National Electricity Market wholesale data showing revenue splits across energy, FCAS, and network support
- PJM Interconnection — Capacity market design documentation and storage participation rules illustrating duration requirements
- UK National Grid ESO — Enhanced Frequency Response tender results and Capacity Market auction clearing prices