Is the warranty's end-of-life capacity the expected life?

No. It is the minimum the manufacturer promises under its conditions. The battery may do better or, outside those conditions, worse; field data from small LFP systems suggests faster ageing than many lab models.

Does peak shaving age a battery less than arbitrage?

Usually it uses fewer full cycles a year, so cycle ageing is lower; calendar ageing continues regardless. Shallow cycles are not automatically harmless, so count them, do not ignore them.

How does temperature matter?

Warranties state their test conditions, often 25 °C, and calendar ageing rises with temperature in the cell models. Ask how the cabinet is cooled and what the warranty says outside its conditions.

Should every business case include a replacement?

Only if the battery reaches its end-of-life capacity inside the project life under the assumed use. If it does, include it; leaving it out can more than double the NPV, as the worked example in the NPV guide shows.

Degradation, cycles and replacement: what the battery warranty tells you

What a warranty actually promises

SolarEdge CSS-OD 197: 8,000 cycles or 10 years, whichever comes first, to 70% state of health (datasheet). The warranty document for its smaller sibling counts a cycle as a full charge and discharge at 95% depth of discharge of the remaining usable capacity, with partial cycles pro rata.
sonnenPro Compact 261: a 10-year product warranty (5 years on the inverter) and 8,000 cycles at 60% state of health (datasheet).
Sungrow ST255CS: a 5-year standard product warranty, extendable to at most 15 years; 65% of nominal energy until 8,760 full cycles at 25 °C and 100% depth of discharge, with a cloud connection and routine maintenance required (Sungrow warranty for C&I ESS).
Fox ESS G-MAX 215: 10 years or 1,492 MWh of discharge throughput to 70%; requires the Fox cloud connection, and assumes two cycles a day without stating a cycle cap (EU warranty).
SMA Storage XL: a 10-year system and capacity warranty, valid only after registration, for 8,500 cycles or 6.5 MWh per kWh (datasheet).

Across Zorowatt's battery database, 23 of the 44 products state an end-of-life capacity: 60% for six, 65% for two, 70% for fourteen and 80% for one home battery; 29 of the 44 state a 10-year warranty.

From warranty to a fade rate

Cycle fade per full cycle = (1 − end-of-life capacity) ÷ warranted cycles. SolarEdge: (1 − 0.70) ÷ 8,000 = 0.00375% per cycle. sonnen: (1 − 0.60) ÷ 8,000 = 0.005%. Sungrow: (1 − 0.65) ÷ 8,760 ≈ 0.0040%.
Equivalent full cycles a year = kWh discharged a year ÷ usable kWh. Illustrative: a peak-shaving battery discharging 90 kWh on 250 working days discharges 22,500 kWh a year; on 245 kWh usable that is 22,500 ÷ 245 ≈ 92 full cycles. A battery that also trades daily may do 250 or more.
Calendar fade: Zorowatt's default is 1% of capacity a year, an estimate; it depends on temperature and on how full the battery is kept.

Three estimates for the same battery

Illustrative: 245 kWh usable, 250 full cycles a year (assumption). Capacity left after 10 years:

From the sonnen warranty terms plus 1% calendar fade: 1% + 250 × 0.005% = 2.25% a year, so 1 − 10 × 0.0225 = 77.5%.
From a cell model: Zorowatt's research ran the BLAST-Lite model for large-format LFP cells (25 °C, half full on average, one full cycle a day at 80% depth and 0.5 C) and got about 77% after 10 years at one full cycle a day, more than the 250 a year here — a research calculation, not a published figure (BLAST-Lite; Gasper et al. 2023).
From field data: LFP home storage systems measured in Germany for up to eight years lost about 2 percentage points of capacity a year (2.2 in capacity tests) at about 200 full cycles a year (Figgener et al., Nature Energy 2024). At that rate, 1 − 10 × 0.022 = 78% to 1 − 10 × 0.02 = 80% after 10 years, at fewer cycles than assumed here.

The three land close together, at about 77–80% after 10 years, but each has a catch. The warranty is a minimum promise under its conditions. The cell model predicts expected life, not warranty life; its documentation says to treat it as qualitative and that pack life is typically 20–30% shorter than cell life. The field data are home systems, and Zorowatt's research notes that their fade of about 2 points a year is higher than many lab models predict. Show the degradation assumption in the report, and test the case with a faster one.

What else the research says

PNNL's 2022 assessment puts the calendar life of LFP at 16 years, and its LFP cycle table gives 2,400 cycles at 80% average depth of discharge, 4,500 at 70%, 8,000 at 60% and 32,000 at 30%, to 60% of rated energy (PNNL 2022).
Shallow cycles are not automatically harmless: a cycle-ageing study of a commercial LFP cell found an unexpectedly strong, partly reversible capacity loss with shallow cycles at medium state of charge (Naumann et al. 2020).

Replacement and the business case

When the battery reaches its end-of-life capacity inside the project life, the case needs a replacement (or an augmentation with new modules). Zorowatt's default replaces it at 60% of the original cost, assuming prices keep falling — an estimate. In the worked example in NPV, IRR and payback for a C&I battery, a replacement in year 14 costs €81,000 and takes a large share of the NPV.
Scaling savings down in proportion to capacity is a simplification. A peak-shaving battery often keeps its full cut until it is too small for the worst day, then loses much of it at once. Zorowatt re-runs the dispatch at a few capacity levels — for example years 1, 5 and 10 — and bills each one.

What to ask the vendor

The end-of-life capacity, and whether it is measured on nominal or usable energy.
The cycle or throughput limit, how a cycle is counted, and which limit applies first.
The conditions: temperature, depth of discharge, C-rate, cloud connection, maintenance, registration.
What the warranty covers: capacity only, or also the inverter, cooling and controls — and for how long each.
What happens at end of life: replacement, augmentation, or a pro-rata payment.

Frequently asked questions

Is the warranty's end-of-life capacity the expected life?: No. It is the minimum the manufacturer promises under its conditions. The battery may do better or, outside those conditions, worse; field data from small LFP systems suggests faster ageing than many lab models.
Does peak shaving age a battery less than arbitrage?: Usually it uses fewer full cycles a year, so cycle ageing is lower; calendar ageing continues regardless. Shallow cycles are not automatically harmless, so count them, do not ignore them.
How does temperature matter?: Warranties state their test conditions, often 25 °C, and calendar ageing rises with temperature in the cell models. Ask how the cabinet is cooled and what the warranty says outside its conditions.
Should every business case include a replacement?: Only if the battery reaches its end-of-life capacity inside the project life under the assumed use. If it does, include it; leaving it out can more than double the NPV, as the worked example in the NPV guide shows.

Related guides

Sources

Last checked: 2026-10-30. Every figure in this guide comes from the source linked next to it, with the retrieval date Zorowatt's research recorded (28–30 October 2026), or from arithmetic shown step by step. Warranty terms as recorded in Zorowatt's battery database (retrieved 28 October 2026); the 10-year capacity figures are illustrative calculations.

Stuck on something else? Ask the community.

Esita küsimus

Uued juhendid otse sinu postkasti

Umbes üks e-kiri kuus: uued juhendid ja videod, lahendused probleemidele, mille kohta inimesed küsivad, ning Zorowatti uudised.

Size it right with Zorowatt

Upload a site's meter data and bill. Zorowatt checks the data, rebuilds the bill, sizes a battery and writes the report. Your first site analysis is free.

Alusta tasuta