Modern grid-scale batteries are reliable machines. What’s risky is everything around them: whether the capacity you paid for shows up, keeps showing up for twenty years, and whether anyone stands behind it when it doesn’t. That risk doesn’t live in the cells — it lives in the contracts you sign in the closing stretch of development. Here’s the map.
The two big structures: wrap vs split
Full EPC wrap. One contractor delivers the whole plant — batteries, PCS, balance of plant — under a single fixed-price, date-certain contract with unified guarantees. You pay a premium for a single throat to choke, and the wrap is only as good as the contractor’s balance sheet.
Split (multi-contract). The owner buys the battery system directly from the OEM under a battery energy storage supply agreement (BESA), contracts balance-of-plant and installation separately, and coordinates the interfaces. This has become the mainstream structure for large projects: DC block prices fell fast enough that owners wanted direct OEM relationships and pricing, and integrators’ wraps couldn’t compete. The cost is that integration risk moves to the owner — when the capacity test fails, the battery OEM points at the PCS supplier, who points at the EPC, and your interface matrix becomes the most important document you own.
Inside the BESA
The clauses that deserve your senior people:
- Scope and interfaces. DC block only? AC block including PCS? Who supplies the EMS, and who integrates it? Every gap in the scope matrix is a change order waiting to happen.
- Delivery and trade risk. Incoterms, title transfer, and — decisive in the US market since 2025 — who bears tariff and duty changes between signing and delivery. Tariff allocation clauses have gone from boilerplate to headline commercial terms.
- Price adjustment. Raw-material indexation clauses were common when lithium prices spiked; with cell prices falling, fixed pricing has returned, but watch for one-way escalators.
- Payment milestones vs security. Prepayments against parent guarantees or letters of credit; retention through commissioning.
Warranties: the heart of the deal
Three distinct promises, often confused:
- Defect warranty. Materials and workmanship, typically a few years. Necessary, not sufficient.
- Capacity / degradation warranty. A year-by-year table of guaranteed retained energy — the commercial embodiment of the degradation curves I covered in the sizing article. Always conditional on an operating envelope: annual throughput or cycles, temperature limits, SoC dwell, C-rate. Operate outside it and the table is void — which is why your trading strategy and your warranty must be designed together, and why BMS data is commercial evidence.
- Availability guarantee. A percentage (high-90s is typical territory) with carefully negotiated exclusions — grid outages, force majeure, owner-caused events — and remedies in liquidated damages or service credits. The definition of “available” (full power? partial? measured how?) is where the negotiation actually happens.
Add to these the serial defect clause — if the same failure appears across some threshold percentage of units, the remedy escalates from repair-as-you-go to fleet-wide campaign at the supplier’s cost. After several well-publicized module recalls in this industry’s history, no serious buyer skips it.
Liquidated damages: pricing failure in advance
- Delay LDs. A daily rate, commonly expressed as a fraction of contract price per day with an overall cap. Worked example: on a $200 million EPC contract with delay LDs of 0.1% per day capped at 10%, you accrue $200,000 per day and exhaust the cap after 100 days. If your revenue at risk is higher than that, you’re self-insuring the difference — knowingly or not.
- Performance LDs (buy-down). If the capacity test comes in short and cure fails, the price adjusts — commonly framed as a $-per-kWh buy-down of the shortfall. The alternative remedy, rejection, is nuclear and rarely exercised.
- Availability LDs. Ongoing operational shortfalls priced against the guarantee.
Caps, exclusive-remedy language, and carve-outs (gross negligence, fraud) determine whether the LD regime is real protection or decoration. Lenders read these clauses first.
The long tail: LTSA and augmentation
The supply deal increasingly comes with a long-term service agreement — preventive maintenance, spare parts, remote monitoring, sometimes an availability guarantee wrapped in — and, more recently, capacity maintenance or augmentation agreements where the OEM commits to keep the plant at contracted energy over time at agreed pricing. These blur the line between procurement and O&M, and they’re a big part of how OEMs defend margin as hardware prices fall.
Bankability: the lender’s checklist
When a lender’s technical advisor reviews your contracts, they’re checking a familiar list: creditworthy counterparties and real parent guarantees; a warranty envelope your revenue model actually fits inside; LD levels that cover debt service during delay; technology with a UL 9540 listing and fleet track record; interface risk allocated to someone who can manage it; and an O&M plan with named parties. That’s bankability in practice — not a stamp, but the absence of unallocated risk.
FAQ
What is a BESA? A battery energy storage supply agreement — the direct contract between a project owner and the battery system OEM covering equipment scope, delivery, price, warranties, and performance guarantees.
What does a BESS degradation warranty cover? A guaranteed minimum retained capacity per year of operation, conditional on staying inside a defined operating envelope of throughput, temperature, SoC, and C-rate.
EPC wrap or split contracting — which is better? Wraps simplify risk but cost more and are only as strong as the wrapper; splits get better pricing and direct OEM warranties but leave integration risk with the owner. Large, experienced owners mostly split; first-timers should think hard before doing so.
The course dissects real contract structures clause by clause — warranty tables, LD math, interface matrices — in the Grid-Scale BESS: Complete Guide.