Battery Buyers’ Contract Gives Seller Claims No Extra Weight

← Back · ACT II · Chapter 9: The Value Control Loop

On October 13, twenty-two organizations, among them utilities, independent power producers, developers, asset owners and an insurer, launched the Energy Storage Contract Alliance, a Swiss non-profit based in Lausanne. Its members represent more than 40 gigawatt-hours of utility-scale battery projects. What they published is a model contract for buying large battery systems, drafted from the buyer’s side. The core clauses are free for anyone to read, buyer or supplier. The technical appendices, where the detailed figures live, are for members.

The alliance explains why it exists in one line on its own site: “By default, the supplier writes your BESS contract.” Its FAQ goes further. Most of these contracts, it says, “start from the supplier’s template, which is built around what the supplier is willing to guarantee rather than what the asset needs to deliver.”

Anyone who has bought an enterprise software platform or an AI service will recognize that template. A quality-assurance consultancy that sells inspection and battery analytics services, and serves as the alliance’s technical secretariat, drafted the first version and also sits on the members’ steering committee, which decides on revisions. Use is voluntary. A buyer can adopt the model whole, take parts of it or change any clause. Even so, the clauses are specific enough to borrow, and a few of them deal with a problem most buyers never put in writing.

Who Owns the Numbers

The clauses I would read first are about evidence. Clause 6.6.3 says “The Buyer owns all data generated by the BESS” and lets the buyer give that data to any third party it nominates, with no consent needed from the seller. Clause 6.6.6 removes any priority the seller’s figures would otherwise carry: “Performance, availability and degradation figures reported by the Seller have no evidential priority over figures derived by the Buyer or the Selected Third-Party from the same data.” The argument can still happen. Disputes over the data go to an independent technical expert under clause 15.2, and under clause 7.3.6 the availability calculation made by the buyer or its nominated third party stands unless the seller gives a reasoned notice of disagreement within twenty business days.

The FAQ is blunter: “A supplier dashboard is not data access.” The contract requires an open interface, sets sampling intervals of a minute or less, and obliges the seller to keep the raw monitoring data for at least ten years after site acceptance. It also deals with the moment most buyers eventually run into, when data goes missing just as performance comes into doubt. Where the gap is the supplier’s doing, the FAQ says the system is then “deemed to have failed its guarantees for that period unless the supplier demonstrates otherwise.” The contract text is more measured. Where the seller is at fault, missing data counts against availability unless the seller can give reliable evidence the system was capable, a capacity figure that cannot be read gets a repeat test at the seller’s expense, and the burden of showing the gap was not its doing sits with the seller. Either way, the buyer is no longer left trying to prove underperformance with records the seller controls.

The Measure Is Set Before Signature

Capacity is measured at commissioning and tested every year after that, against a degradation curve written into the agreement. The model’s default floor is 80 percent of usable capacity at the end of the warranty. In the fourteen days before a performance test, the seller may not perform cell balancing, recalibrate its state of charge or change its firmware. A status light does not count as performance either: “A reported ‘online’ or ‘ready’ status does not evidence availability; only the measured capability to deliver Contracted Performance does.”

The model also sets, in advance, how a shortfall is priced, point by point for lost efficiency or availability. For performance, clause 7.2.5 says the damages are calculated on the shortfall and its impact on the buyer’s expected revenue, by a method set out in a members-only appendix. The total is capped, but paying it does not end the obligation to restore performance unless the buyer agrees otherwise. The alliance’s reasoning applies well beyond batteries. A cap with no carve-out, it says, “lets the supplier treat LDs as a price list and buy its way out of ever fixing the system.”

None of this is exotic. The model asks the buyer to write down, before signature, what the asset has to deliver, how it will be measured, who keeps the record and what a miss will cost, while the buyer still has the leverage to get those terms.

What a Good Supplier Gets Out of It

It would be easy to read this as buyers lining up against manufacturers. The alliance’s press release makes a different case, and a supplier that delivers should welcome it. A public baseline, it says, “lets suppliers respond to tenders against known requirements, and lets smaller developers negotiate on the same footing as large utilities.” The contract sets no prices, and the alliance takes no part in its members’ commercial negotiations.

Pedro Garcia Hrdy of MN8 Energy, a member of the steering committee, describes the cost of the old way. Negotiating guarantees, testing and risk allocation from scratch on every project “adds cost, delays financing, and creates uncertainty for developers, operators, offtakers, lenders, and OEMs alike.” Common terms, he says, will help “make risk easier to price.”

A manufacturer whose batteries hold their capacity has little to fear from clause 6.6.6. It gains buyers who can show the difference between its product and one that falls short.

Carrying It Over to Software and AI

A battery is easier to measure than a CRM rollout or an AI agent. Usable capacity is a physical quantity. The value of a software purchase has to be defined before anyone can measure it, and the vendor often offers to do the defining, along with the dashboard that reports on it afterward.

That makes the battery buyers’ order of operations useful to a software buyer too. Their model sets the measure, the baseline and the data rights before signature, and it gives the seller’s reports no priority over the buyer’s own figures. The alliance’s FAQ has a line for what happens otherwise: “The gap costs nothing at signature and everything at commissioning.” For a software or AI purchase, that bill often arrives at renewal.

A single buyer can do the same thing for a single contract. Write down the result the purchase is meant to produce, the baseline it will be measured from, the data the buyer will own, and what a shortfall costs, and send it to every vendor before award so each one answers the same document. That document is a Request for Value. The battery contract’s clause on whose figures count is a sound model for the part of it that covers data.

Sources: Energy Storage Contract Alliance, press release, October 13, 2026 (https://escabattery.org/downloads/press/ESCA-Press-Release-2026-10-13.pdf); ESCA site and FAQ (https://escabattery.org/); ESCA Model BESS Supply & Purchase Agreement, clauses 6.6, 7.2, 7.3, 14.1 and 15.2 (https://escabattery.org/).

Keep reading: ACT II · Chapter 9: The Value Control Loop