Solar panel degradation and the 25-year warranty: modelling year one and year twenty-five honestly

Key Takeaways

• A 25-year performance warranty is a floor, not a forecast: LONGi's Hi-MO 6 guarantees 86.8% of nameplate output at year 25, Trina's Vertex S+ 87.4%, and premium n-type panels from REC and SunPower/Maxeon 92%.
• Averaged over 25 years, those floors imply roughly 0.3-0.53% annual degradation, not the round 0.5% or 0.8% figures often used as shorthand.
• Field studies have found real-world fleets degrading faster than warranty curves suggest, at a median of around 0.8-1.09% a year, which matters more than the headline warranty number for a payback model.
• Modelling flat output for 25 years overstates lifetime generation by roughly 6-10% on a typical rooftop system, enough to shift payback by more than a year.
• The warranty protects you if output falls below the contractual curve — it does nothing if your model never checked what that curve actually implies.

A 25-year performance warranty does not promise 25 years of unchanged output. It promises a floor: a guaranteed minimum percentage of nameplate capacity at year 25, and that percentage sits well short of 100%. LONGi's Hi-MO 6 warranty floor is 86.8% of nameplate output at year 25, and Trina's Vertex S+ sits at 87.4% (Clean Energy Reviews, Best Solar Panels 2025, retrieved 2026-09-07) — run your own roof through the solar payback calculator with that curve built in rather than a flat output assumption, and the payback period moves by more than a rounding error.

Most payback models still treat year-one output as the number that repeats for a quarter of a century. It is the most common shortcut in solar financing, and the one that most consistently overstates the return, because the gap between "flat output" and "warranty-curve output" on a typical system runs into hundreds of thousands of kilowatt-hours over 25 years.

The warranty numbers are not uniform either. Premium n-type cells from REC and SunPower/Maxeon guarantee 92% of nameplate output at year 25, against 86.8-87.4% for mainstream monocrystalline PERC and TOPCon panels from LONGi and Trina (Clean Energy Reviews, Best Solar Panels 2025, retrieved 2026-09-07; REC Group warranty documentation, retrieved 2026-09-07). That spread is worth pricing into a quote comparison, because it changes the output curve the payback calculation should assume, not just the price per watt.

The warranty floor is a single checkpoint, not a curve you can read off directly

A performance warranty states one guaranteed number at year 25, sometimes a second at year 10 or 12. It does not hand you the annual degradation rate — you have to back it out. LONGi's 86.8% floor and Trina's 87.4% floor both imply a total loss of roughly 12.6-13.2 percentage points over 25 years, which averages to about 0.5% a year treated as linear (Clean Energy Reviews, Best Solar Panels 2025, retrieved 2026-09-07). REC's and SunPower/Maxeon's 92% floors imply a shallower average of roughly 0.32% a year — the gap between an n-type heterojunction or IBC cell and a mainstream PERC or TOPCon cell shows up here, not in the datasheet's efficiency percentage.

That 0.3-0.53% band is a modelling convenience backed out from the contract, not a guarantee of smooth, even decline year by year.

Year one is not like every year after it

Degradation is not linear from day one. Light-induced degradation (LID) — the loss that happens in the first months of exposure as the cell structure settles — used to make year one disproportionately worse than any year that followed. High-efficiency n-type panels now show LID "as low as 0.25% of power loss per year" (Clean Energy Reviews, Most Efficient Solar Panels, retrieved 2026-09-07), a meaningfully narrower gap than older p-type designs carried.

The practical consequence: a straight-line average from 100% to the year-25 floor slightly understates early output and overstates late output, but the error mostly cancels out over 25 years — defensible for a payback model, less so for a five-year exit or lease buyout, where the first few years' shape matters more than the 25-year average.

Field data does not always sit inside the warranty curve

The warranty curve is what the manufacturer will honour; it is not necessarily what a given roof experiences. Independent testing exists because installed-fleet performance and datasheet claims can diverge — Kiwa PVEL has evaluated more than 600 module types from over 75 manufacturers on exactly this gap, feeding its annual PV Module Reliability Scorecard (Kiwa PVEL, Solar Module Testing and Reliability, retrieved 2026-09-07). Broader field studies of installed systems have found median annual degradation running at roughly 0.8% for non-residential installations and 1.09% for residential ones, both above the 0.5% figure often used as a rule of thumb (Wikipedia, Photovoltaic degradation, retrieved 2026-09-07).

Run that 0.8% figure to year 25 and output has fallen to 80% of nameplate — below LONGi's own 86.8% contractual floor. A real installation tracking that field average, rather than the warranty curve, would be sitting in breach-of-warranty territory by year 25, provided someone had been recording output against the datasheet curve well enough to prove it.

Building the curve into a payback model, not just the price

Take a 200 kWp rooftop array generating 340,000 kWh in its first full year. Modelled flat for 25 years, that is 8,500,000 kWh of lifetime generation. Apply LONGi's own warranty curve — a linear decline from 100% to 86.8% — and the 25-year average output falls to 93.4% of year-one generation, bringing lifetime output down to roughly 7,940,000 kWh: about 560,000 kWh, or 6.6%, less than the flat model assumed.

Model the same array against the field-study non-residential average of 0.8% a year instead, and average output drops to 90% of year-one generation — roughly 7,650,000 kWh, close to 10% below the flat assumption and below what the manufacturer's own warranty curve implies. That is the gap between what the datasheet promises and what a comparable fleet actually delivered, and a payback case built only on the first number is the one most likely to disappoint whoever signed off on it. Neither adjustment is exotic — both are a single extra input line in a spreadsheet, or a toggle in a calculator that already asks for system size and tariff.

Reading the datasheet before the quote gets signed

A supplier's headline efficiency percentage describes the panel at commissioning, not at year 20. The number worth asking for is the warranty's year-25 floor and, ideally, a year-10 or year-12 checkpoint, because that tells you whether the decline is a smooth straight line or a steeper early step followed by a flatter tail. Two panels quoted at the same price per watt can carry meaningfully different 25-year output, which is why the degradation curve belongs in the solar panel comparison at quote stage, not as a footnote read after the contract is signed.

Asking for the degradation assumption also gives a baseline to test against later. If output is ever disputed against a warranty claim, the comparison has to be made against the manufacturer's stated curve under standard test conditions, not a vague sense that the system "seems to be producing less than it used to."

Frequently asked questions

Does degradation reduce my savings by the same amount every year?

No. The loss is smaller in early years and larger in absolute kilowatt-hours later, since the percentage decline applies to a base that has already shrunk. Real panels can also behave differently in year one due to light-induced degradation, though modern n-type cells have narrowed that effect.

Should a payback model use the warranty floor or the field-study average?

Use the warranty floor as the contractual worst case and the field-study average as the realistic case, and run both. If payback only clears the hurdle rate under the optimistic flat-output assumption, flag that before the contract is signed.

What happens if my system underperforms the warranty curve?

You have grounds for a claim, but the manufacturer will want measured output compared against the datasheet's stated curve, usually via independent test — only possible if generation has been logged consistently from commissioning.

Do all manufacturers structure their warranty the same way?

No. Some publish a single year-25 floor and let the buyer infer a linear rate; others state a steeper early step followed by a flatter rate. Read the actual curve rather than assuming a round number applies.

The bottom line

The 25-year warranty is not a performance forecast, and treating it as one is the most common way solar economics get overstated. The number that matters is the year-25 floor in the manufacturer's own document, converted into an annual rate, checked against what comparable installed fleets have actually delivered, and run through the calculator alongside — not instead of — the flat-output case.

Whichever curve is used, the decision it should change is not whether to install solar, but which panel, at which price, clears the reader's actual hurdle rate once the honest output curve replaces the optimistic one.

Figures were verified on 7 September 2026 against Clean Energy Reviews' manufacturer warranty comparison, REC Group's published warranty documentation, Kiwa PVEL's testing programme summary, and Wikipedia's Photovoltaic degradation article. WebSearch was unavailable for this session (budget exhausted), so verification relied on direct WebFetch retrieval against these sources rather than the usual NREL and manufacturer-datasheet mix; several primary sources (nrel.gov, individual Jinko/LONGi/Trina PDF datasheets) could not be reached and were excluded rather than estimated.