Key takeaways
- Energy storage as a service (ESaaS) is a contract in which the customer pays for a defined storage outcome, such as available capacity, a verified saving or guaranteed uptime, while the provider keeps control of the battery and carries the performance risk.
- Industrial energy storage providers contract outcomes through three archetypes: availability, savings and infrastructure outcomes, and performance guarantees.
- Providers moving from equipment sales to outcome-based contracts report deal values two to three times higher than equivalent CAPEX transactions, with stronger retention and renewal economics.
- Choose the archetype whose outcome you can measure continuously, price against field conditions rather than best case, and finance on a tenor that matches the contract.
- Providers who skip the governance and verification layers to accelerate deployment consistently encounter billing disputes, margin leakage and contract renegotiation within 18 months.
Battery manufacturers and integrators keep hearing the same instruction from their boards: build a service offer. Very few are told which one.
That gap matters, because industrial markets are not moving toward subscription in general. They are moving toward contracted outcomes, and only some outcomes can be contracted. This guide sets out the three contract archetypes that work in battery energy storage (BESS), what each one charges for, the data it depends on, where it leaks margin, and a rule for choosing between them. It draws on P2S research across more than 500 industrial As-a-Service cases in 12 industries, summarised in our whitepaper Beyond the Meter.
What energy storage as a service actually means
Definition. Energy storage as a service (ESaaS) is a commercial model in which a provider delivers battery storage as a contracted outcome (usable capacity, uptime, a verified economic result or a guaranteed performance level) and is paid per month, per kWh, per MW made available or per result, instead of selling the battery as a one-off capital purchase. The provider keeps ownership or operational control of the asset and absorbs lifecycle risks such as degradation, maintenance and obsolescence.
Put simply, battery energy storage as a service turns the BESS business model from a one-off sale into recurring revenue streams tied to what the battery delivers. The model applies to BESS manufacturers and integrators, energy service providers, commercial and industrial (C&I) sites, virtual power plant (VPP) operators, DSOs and TSOs, and regulated asset owners buying flexibility instead of network reinforcement. The commercial logic holds across all of them. The contract mechanics do not, so this article is written from the provider's side of the table.
ESaaS is not a lease with a new label. A lease changes when the customer pays, but the customer still carries the asset risk. A power purchase agreement prices delivered energy per unit. In ESaaS the provider sells a result and is accountable when it is not delivered.
Why hardware margin is moving to contracts
Four structural signals are pushing margin away from the cell and toward whoever can measure, contract and operate the outcome.
| Signal | Data point | Source | What it means for providers |
|---|---|---|---|
| Hardware is commoditising | Li-ion pack prices fell 20% year on year to $115/kWh in 2024. In 2025, stationary storage packs reached about $70/kWh and the benchmark cost of a four-hour battery project fell 27% to $78/MWh | BloombergNEF | The margin premium has to come from services, software and guarantees |
| Grid connection is a bottleneck | About 890 GW of storage sat in US interconnection queues at end-2024, with median queue duration above four years. At least 16 EU Member States face connection queues or related grid bottlenecks | Lawrence Berkeley National Laboratory; European Commission | Service-led and behind-the-meter models win on speed, not only on economics |
| BESS is now core infrastructure | Global battery storage deployment more than doubled in 2023 to over 90 GWh a year, with more than 190 GWh installed | IEA, 2024 | Buyers apply mission-critical procurement logic: guaranteed performance, one accountable party, predictable cost |
| Buyers want outcomes, not ownership | 18.8 billion connected IoT devices by end-2024 | IoT Analytics | Continuous verification is mainstream, so the constraint is commercial, not technical |
The strategic risk for a hardware provider is not missing growth. It is scaling volume while margin migrates into software, lifecycle services, financing and performance accountability. Margin in energy storage service models comes from risk transfer, lifecycle monetisation, software-enabled dispatch, switching costs and capital efficiency for the buyer, not from the cell.
Industrial buyers have seen this before. Compressed air, which Atlas Copco sells per normal cubic metre (Nm³) of guaranteed air, became a contracted utility once measurement and governance matured. Storage is on the same trajectory, only faster.
The three archetypes at a glance
| Availability | Savings & infrastructure outcomes | Performance guarantees | |
|---|---|---|---|
| Customer pays for | Readiness: usable capacity or uptime within defined conditions | A defined economic or operating result | Guaranteed operating performance under defined conditions |
| Typical value metric | Per battery per month, per usable kWh, per operating hour or cycle band, contracted MW in service windows | Per site per month, per contracted kW or kWh, resilience fee, shared savings, deferred network investment value | Availability % by service window, response or restoration time, power-quality thresholds, efficiency floors |
| Best for | Fleets, material handling, backup systems, stationary BESS, grid operators procuring firm flexibility | C&I sites, campuses, transit depots, utilities and DSOs seeking non-wires alternatives | Data centres, hospitals, critical infrastructure, utilities, TSOs |
| Key data requirement | Real-time state-of-health monitoring | A clean baseline, or dispatch rights and settlement rules | Telemetry both parties can verify independently |
| Core provider risk | Degradation priced at best case | Shared savings and capacity access mixed in one contract | Insurance-level liability priced as a maintenance contract |
Archetype 1: availability
Definition. An availability model charges the customer for readiness rather than ownership: usable capacity, uptime or availability within defined service conditions. The provider owns the lifecycle and residual value risk.
For a single site, availability means paying for usable capacity or uptime within agreed conditions. For grid operators and aggregators it means storage as a grid service: they contract available distributed capacity for dispatch during defined windows, instead of owning the assets or building a peaker plant or substation.
The buyer is protected from residual value risk. Batteries reach end of life at approximately 80% state of health, and in an availability model the provider absorbs that decline.
What we see in the field. Battery-as-a-service practitioners in material handling told us their best availability offers use cell-phone-style pricing: a fixed monthly energy package with defined usage parameters, not pure pay-per-kWh billing. Customers prefer predictability, and providers get fewer billing disputes. One practitioner called the switch from variable kWh billing to a fixed monthly plan the single most important commercial design decision they made.
Connectivity is not optional. The IoT layer does three jobs at once: accurate billing, enforcement of the guarantee, and contractual protection if the customer misuses the asset. A CFO we interviewed called it non-negotiable, because monitoring usage protects residual value, and residual value drives the financing economics. See why IoT is a prerequisite for XaaS models and value metrics and metering.
Public examples follow the same shape. Hoppecke offers rental, financing and usage-based billing with remote maintenance, positioned on lower upfront investment and predictable monthly charges. UgoWork runs a pay-per-use energy model for forklift fleets with 24/7 remote monitoring and funds the hardware through a dedicated financial partner, so it can grow the software offer without putting its balance sheet behind every deal. We unpack that structure in our UgoWork podcast episode.
Where it breaks. Margin leakage in battery availability models rarely starts with discounting. It starts with degradation assumptions that were never stress-tested against real operating conditions.
Archetype 2: savings and infrastructure outcomes
Definition. A savings and infrastructure outcome model charges the customer for a defined operating or economic result, such as resilience, deferred infrastructure investment or validated energy cost savings, rather than for the asset that produces it.
For C&I buyers this archetype removes the capital barrier and hands operational complexity to the provider. For utilities and DSOs it can aggregate decentralised storage into a virtual power plant that defers central network investment. This is the logic of non-wires alternatives: the network operator contracts flexible distributed capacity instead of reinforcing a substation. It buys the same kind of guaranteed utility a factory buys when it stops owning compressors and contracts air by volume.
The prerequisites are strict. Clean dispatch rights, verifiable availability windows and settlement rules must be agreed before signing.
Two public cases show it working. In Montgomery County, Maryland, AlphaStruxure designed, financed, built and operates a 6.5 MW microgrid for the Brookville Smart Energy Bus Depot. The county electrifies 70 buses with no upfront cost and avoids an interconnection delay that would have stalled the project by years. The county's only obligation is to use the service. Stem pairs storage with AI-driven optimisation on its PowerTrack platform to deliver demand charge reductions and verified savings for around 16,000 customers in more than 50 countries. In both cases the provider controls the levers that create the outcome, and that control is what makes the outcome billable.
What we see in the field. A cooling-as-a-service provider we studied started with shared savings and deliberately abandoned it after disputes over actual versus projected savings. It moved to a consumption model in which the customer pays for verified output. The transition took years, a proprietary metering system, a dedicated asset management division and several rewrites of the contract template.
Where it breaks. Mixing shared savings logic with capacity access logic in the same contract creates irresolvable billing disputes, because one depends on a verified baseline and the other on dispatch control and settlement discipline. And never promise savings without a clean baseline and a credible verification method agreed before signing. If you cannot prove the savings, you cannot bill for them.
Archetype 3: performance guarantees
Definition. A performance guarantee model charges the customer for guaranteed operating performance under defined conditions, with financial remedies such as liquidated damages when the provider falls short. It is the highest-value archetype and the most demanding one to contract.
Here the provider sells downside risk transfer. A data centre or hospital buys less operational uncertainty and one accountable party for performance across the contract life. Maintenance is simply how the provider delivers it.
That makes the contract an insurance-like promise, and it must be priced like one. Force majeure and site-readiness exclusions have to be explicit and agreed before signing, or the provider pays penalties for grid failures it did not cause. Our guide to risk allocation and guarantees covers the wider mechanics.
A P2S case: critical-power BESS. A European provider of mission-critical power systems signed a performance guarantee with a large financial services campus that needed uninterruptible backup power and peak-load support. Three design decisions made it viable:
- Explicit exclusions. Third-party grid events outside the provider's control were defined before signing, limiting liability to failures within its operational scope.
- A shared source of truth. Telemetry both parties could verify independently was a contractual requirement, not a technical afterthought.
- A priced risk premium. The monthly fee carried a premium for the downside liability, and internal modelling showed it cost less than the expected emergency call-outs and reactive replacements under a conventional service agreement.
At industrial scale, Wärtsilä's Guaranteed Asset Performance model applies the same logic to power generation: guaranteed reliability, availability and efficiency, with defined remedies for non-performance. Its lifecycle agreements cover more than 30 GW of capacity at more than 700 installations. That scale lowers the unit cost of carrying the risk. Smaller providers have to replicate it through partnerships or operational depth.
Where it breaks. If exclusions, causality rules and remedy caps are vague, the provider has priced a maintenance contract while absorbing insurance-level liability.
Go deeper. The full archetype tables, the CEO watch-outs for each, and the six margin risks we see repeat across more than 500 industrial As-a-Service cases are in the Beyond the Meter whitepaper. Download Beyond the Meter
How to choose: a decision rule
Choose the archetype whose outcome you can measure continuously, price against field conditions rather than best case, and finance on a tenor that matches the contract. That gives three tests, applied in order.
- Can you measure it? The outcome must be observable continuously, from data both parties accept as authoritative. If not, no archetype works yet. Build the telemetry first.
- Can you price the risk at field conditions? Degradation, augmentation, dispatch and misuse must be modelled on how the asset will actually run, not on the datasheet.
- Does the financing fit? Contract duration has to match debt tenors and the asset's depreciation schedule. Self-fund the first three to five projects to build the track record lenders require. A special purpose vehicle becomes viable at roughly $30 to $50 million of deployed assets. Expect cumulative outcome-based revenue to overtake the equivalent CAPEX stream after four to seven years, the curve we explain in the Fish Model and in financing and cash-flow structuring.
Then match the archetype to the buyer's real problem:
| If the buyer's main problem is | Start with | You must already have |
|---|---|---|
| Upfront cost, and residual value risk on a battery they do not want to own | Availability | State-of-health telemetry and a degradation model tested against field data |
| A capital budget it cannot unlock, a grid connection it cannot wait for, or network reinforcement it wants to defer | Savings & infrastructure outcomes | A verified baseline, or agreed dispatch rights and settlement rules, never both in one contract |
| Downtime with immediate economic consequences | Performance guarantee | Explicit exclusions, independently verifiable telemetry, a priced risk premium, and the scale or partners to pool the risk |
A provider can run several archetypes across its portfolio. What it cannot do is let two conflicting logics share one contract.
Six margin risks that surface by month 18
Recurring billing does not create margin by itself. When providers skip governance and verification to ship faster, the same six risks appear, usually within 18 months.
- Underpriced technical risk. Degradation, augmentation and dispatch are modelled at best case. Mid-life BESS augmentation, typically needed in years 8 to 10 to hold contracted capacity, is the most commonly underpriced cost in long-term storage contracts.
- Weak contract boundaries. Outcome language without strict scope and force majeure exclusions creates open-ended liability. Every unclear boundary eventually becomes a billing dispute.
- Poor telemetry quality. In energy storage, disputes about performance are almost always disputes about whose data is authoritative.
- Manual billing. Manual processes cannot handle large fleets of variable monthly invoices without significant headcount, which erases the scale benefit.
- Financing mismatches. Contract durations that do not match debt tenors or depreciation schedules create cash-flow gaps and balance sheet pressure.
- Misaligned sales incentives. Sales teams undercut the service offer with a CAPEX alternative to hit booking targets. Commission plans have to be redesigned alongside the commercial model.
Hoppecke, UgoWork, AlphaStruxure, Stem, Atlas Copco and Wärtsilä differ in scale, geography and segment, yet each combines five capabilities: they measure with auditable telemetry, contract with clear exclusions and capped liability, deliver across the full chain, finance with capital matched to contract tenor, and scale with repeatable contracts and aligned incentives. Skip one and margin, retention or scale breaks.
Frequently asked questions
How does battery energy storage make money as a service?
BESS revenue streams in a service model are recurring fees instead of a one-off sale: fees for available capacity or uptime (per battery per month, per usable kWh, or per MW available in grid service windows), fees tied to an economic result (per site, resilience fees, shared savings or deferred network investment), or fees for guaranteed performance backed by liquidated damages.
What is the difference between energy storage as a service and a PPA?
A power purchase agreement sells delivered energy at a price per unit. Energy storage as a service sells a contracted outcome, such as available capacity, a verified saving or guaranteed uptime, and transfers operational, maintenance, degradation and obsolescence risk to the provider.
Who owns the battery in an energy storage as a service contract?
Usually the provider, its financing partner or a special purpose vehicle, not the customer. The customer contracts the outcome and the asset holder carries the residual value risk. UgoWork funds its batteries through a dedicated financial partner, and AlphaStruxure designs, finances, builds and operates the microgrid it contracts to Montgomery County.
What happens when the battery reaches 80% state of health?
Batteries are generally treated as reaching end of life at approximately 80% state of health. In an availability contract that decline is the provider's problem, managed through augmentation, typically in years 8 to 10 of a long-term contract, and through second-life and recycling pathways that become manageable at fleet scale.
Is energy storage as a service more profitable than selling the equipment?
It can be. P2S research across more than 500 industrial As-a-Service cases finds deal values two to three times higher than equivalent CAPEX transactions, with stronger retention and renewal economics. That only holds if degradation, liability and financing are priced correctly, and the cash crossover typically takes four to seven years.
Build the contract before you build the offer
The firms winning in energy storage services do not have the cheapest cells. They decided early which outcome they could prove, priced the risk they took on, and financed it on the right tenor. If your board has asked for a service offer, answer those three questions first. The archetype follows.
Beyond the Meter: how Energy Storage & Utilities win with outcome-based service models. The full archetype tables, six practitioner cases, the CEO decision checklist and the financing appendix, drawn from more than 500 industrial As-a-Service cases. Download the whitepaper or book a conversation with a partner.
Related reading
- Servitization for industrial OEMs
- Outcome-based contracting and SLAs
- Servitization in HVAC with Kaer: from selling chillers to selling cooling
- Energy Storage & Utilities at P2S
Sources: BloombergNEF Battery Price Survey 2024 and 2025 benchmark; Lawrence Berkeley National Laboratory; European Commission; IEA (2024); IoT Analytics; P2S Management Consulting research across 500+ industrial As-a-Service cases. Practitioner cases are anonymised except where publicly disclosed.