Merchant vs. contracted

What it is (precise)

In a contracted structure an off-taker pays a fixed availability or capacity fee — in $/kW-month or $/kW-year — and usually controls dispatch under a tolling agreement, or buys a defined capacity product over a 10-20 year term. Revenue is decoupled from hourly prices: the off-taker absorbs market risk and keeps the trading upside.

The contract names the physical deliverables — MW, usable MWh at the point of interconnection, round-trip efficiency, availability, response time — plus a capacity-retention schedule the asset must hold for the full term. You first meet this structure as a signed tolling agreement or PPA, and every number written into it becomes a hard engineering obligation.

In a merchant structure the project bids its own capacity into the wholesale energy, Ancillary services and capacity markets run by the local ISO / RTO, earning whatever those markets clear at, hour by hour.

Revenue is the sum of stacked value streams — Energy arbitrage, Frequency regulation, reserves and capacity payments — optimized in near-real time by the energy management system and a route-to-market provider; this Revenue stacking is what makes the merchant case pencil at all. There is no fixed floor: a summer of scarcity pricing can be extraordinarily lucrative, while a mild, low-volatility year can leave the asset well short of plan and short of its debt service.

Why it matters in a real grid-scale project

The merchant/contracted choice is the single biggest driver of how a stationary BESS is financed. A contracted project with a creditworthy off-taker supports non-recourse project debt at high gearing and a low cost of capital, because the debt service coverage ratio is computed against a secured cash flow.

A fully merchant project is financed against a probabilistic revenue forecast, so lenders advance less, demand higher returns and size a larger equity cushion — if they lend at all. Most real deals therefore split the difference: a contracted floor or partial toll makes the senior debt bankable, while a merchant tail is retained for the upside.

The choice reaches straight into engineering and O&M, which is where a student first sees revenue strategy collide with hardware limits. A merchant asset chasing arbitrage and fast Ancillary services tends to cycle harder and at deeper depth of discharge, accelerating LFP capacity fade and pulling augmentation forward; a tolled asset cycles to a schedule the off-taker controls, making degradation more predictable, though it is still bound to contracted capacity-retention and availability targets.

Either way the cell warranty, long-term service agreement and augmentation plan must match the expected cycling regime — a warranty written for one cycle per day will not survive a merchant strategy that cycles twice daily for ancillary services.

Typical values and standards

Contracted terms commonly run 10-20 years — think CAISO resource adequacy — with shorter 5-10 year tolls seen in ERCOT; availability guarantees typically sit at 95-98% (up to ~99%), round-trip efficiency commitments around 85-90% AC-AC at the point of interconnection for modern LFP systems, and cycling is frequently capped near one full cycle per day or an equivalent annual MWh throughput.

Publicly reported tolling and capacity prices for 4-hour systems have varied widely — very roughly $8-20/kW-month depending on market, vintage and term — so treat any single number as a snapshot, not a benchmark. Most systems today, contracted or merchant, run 2-4 hour duration, which is a 0.25C to 0.5C rate.

Merchant revenue is modeled as a distribution, not a single number: independent consultants issue P50 and P90 forecasts, and lenders size debt against the downside case with material headroom. As rough guides, contracted cash flows are often underwritten near a 1.30-1.40x minimum DSCR while merchant cash flows demand on the order of 1.75-2.00x or more, which is exactly why the same asset supports far less debt as a merchant.

In the US the Investment Tax Credit — 30% base for standalone storage under current law, with bonus adders — reduces the capital at risk under either model but does not change how market risk is allocated between the project and its off-taker.

Safety and code compliance are identical across both models and are a precondition to any financing: the system must be UL 9540 listed, with thermal-runaway propagation behavior characterized by the UL 9540A test method, installed per NFPA 855 (which invokes deflagration protection under NFPA 68/69), behind a grid-code-compliant interconnection.

Chemistry interacts with the revenue model mainly through cycling: LFP dominates stationary storage for its higher thermal-runaway onset temperature and longer cycle life, which suits hard-cycling merchant duty better than higher-energy-density NMC. None of these certifications changes with the revenue model, but the cycling the model implies is what stresses them.

How it shows up in specs, studies and contracts

On the contracted side the tolling or capacity agreement is where the engineering meets the money, so read it like a spec sheet. Check the definitions clause first: are the contracted MW and MWh measured at the point of interconnection, net of PCS, transformer and auxiliary losses, or at the DC terminals?

Then verify the capacity test protocol and its frequency (usually an annual full-discharge test), the capacity-retention curve and augmentation obligation, the availability guarantee with its liquidated damages, the round-trip efficiency guarantee and its measurement basis, and the cycling or throughput cap. Reconcile every one of these against the cell warranty and the vendor's degradation model before signing; that reconciliation is the binding constraint.

On the merchant side the working documents are the market consultant's P50/P90 revenue forecast, the route-to-market or optimization agreement (fixed fee, revenue share, or a floor-plus-share product), and the ISO / RTO market rules governing qualification, state-of-charge management and performance measurement.

The vendor question to ask is whether the dispatch strategy the optimizer plans to run fits inside the warranty's cycle and throughput limits, and who holds the bidding keys — the owner, the optimizer, or a hybrid. The physical constraint still governs: the interconnection agreement caps what any strategy can deliver, since no bid can exceed the studied injection and withdrawal limits at the POI.

Common pitfalls

The classic trap is contracting numbers on the wrong basis. A 100 MW / 400 MWh nameplate system is not a 100 MW / 400 MWh contracted system: usable energy at the POI is net of the usable SOC window, round-trip losses and auxiliary consumption, and it declines from BOL toward EOL. A toll that guarantees beginning-of-life energy for 20 years without a costed augmentation plan quietly transfers the entire degradation risk to the owner. Similarly, an RTE guarantee is meaningless until the document states AC-AC versus DC-DC and the ambient and cycling conditions under which it is measured.

On the merchant side the recurring failure is extrapolating today's prices. Ancillary services markets are shallow: Frequency regulation and fast-response products saturate quickly as storage capacity builds out, which is exactly what happened in early storage markets, and long-run merchant cases lean increasingly on Energy arbitrage spreads instead.

A revenue model that assumes current ancillary clearing prices persist for 20 years will not survive lender diligence. Finally, hybrid deals need aligned incentives: an optimizer paid a share of gross revenue has little reason to protect cycle life unless the agreement makes the warranty's cycle and throughput limits a hard dispatch constraint.