Battery warranties are more complex than module warranties and considerably more consequential, because degradation is faster and the operating regime affects it directly.
Calendar ageing happens regardless of use. Chemical processes inside the cell progress with time, accelerated by temperature and by sitting at high state of charge. A battery installed and never cycled still loses capacity.
Cycle ageing happens through use. Each charge-discharge cycle causes small irreversible changes. Deeper cycles cause more damage than shallow ones, and faster rates cause more than gentle ones.
Both run simultaneously. Which dominates depends on your application: a backup battery that sits idle degrades mostly by calendar, an arbitrage battery cycling daily degrades mostly by use.
A typical commercial LFP warranty says something like: 10 years, or 4,000 cycles, or 12 GWh throughput – whichever comes first – with at least 70% of original usable capacity retained.
Read each limb, because whichever binds first ends the warranty:
This is where the real content sits, and it is usually in an annexe:
Temperature limits. Almost every warranty specifies an operating temperature band, often something like 0-35°C for sustained operation. Exceeding it repeatedly can void cover. In practice this means the thermal management system is a warranty-critical component, not an optional extra – and in hot climates, active cooling is usually mandatory.
State of charge limits. Holding a lithium battery at 100% for extended periods accelerates calendar ageing significantly. Some warranties limit how long the system may sit above a given SoC. This directly conflicts with backup applications, which want high SoC always – a genuine tension worth raising with the supplier before purchase.
Charge and discharge rate limits. Expressed as a C-rate. Exceeding it, even briefly and even if the hardware permits, can be excluded.
Monitoring and data obligations. Many warranties require the system to remain connected so the manufacturer can log operating data. Disconnect it for a year and you may have no claim.
| Regime | Typical annual capacity loss (LFP) |
|---|---|
| Standby, good thermal management, moderate SoC | 1.5 - 2.5% |
| One cycle per day, 80% DoD, temperature controlled | 2 - 3% |
| Two cycles per day, deep discharge | 3.5 - 5% |
| Poor thermal management, hot climate | 4 - 7% |
Degradation is not linear – it is typically faster in the first year, then settles, then accelerates again near end of life. Warranty curves usually reflect this, and a straight-line assumption in a financial model will be wrong in both directions at different times.
Give them your intended duty cycle – cycles per day, depth, ambient conditions – and ask them to state the expected usable capacity at year 5 and year 10 under that regime, and confirm it stays inside warranty terms.
A supplier who can produce that in writing has modelled your application. One who quotes a generic degradation curve has not, and the gap between those two positions is where disappointing outcomes come from.