How Long Do LED Grow Lights Last? L70, LM-80 and TM-21 Explained
A horticultural LED fixture rarely stops working. It quietly delivers less light than it did when it was installed, and the crop plan that was sized on day one slowly stops matching the light that is actually reaching the canopy. The numbers printed on a datasheet — L70, LM-80, TM-21 — describe that decline, but they describe it under rules that are frequently misread. This article explains what each term means, why the projection rules cap the claims that can honestly be made, what the published data actually shows for horticultural luminaires, and how to turn hours into a replacement decision.
- What "lifetime" means for an LED fixture
- The 6X rule: why a 100,000 hour claim is unsupported
- What horticultural luminaires actually report
- Turning rated hours into calendar years
- What the standards do and do not cover
- Planning replacement around PPFD, not hours
- Frequently asked questions
- About the publisher
- References
What "lifetime" means for an LED fixture
The first thing to unlearn is the idea that an LED has a service life in the way an incandescent lamp does. An LED does not contain a filament that snaps; its light output declines steadily with operating hours, so the industry expresses life as the projected number of hours until output reaches a stated fraction of its initial value. The Illuminating Engineering Society uses this construction when it discusses claims made "to L70", where L70 denotes the projected operating hours at which luminous flux falls to 70 percent of its initial level; L80 and L90 follow the same convention at 80 and 90 percent.[1]
The measurement and the projection are two separate steps, and they use two different documents. A fixture's performance claims rest on a test method and a projection method, and it is worth knowing which is which before comparing two datasheets.
| Term | What it is | What it is not |
|---|---|---|
| L70 / L80 / L90 | The projected operating hours at which output reaches 70, 80 or 90 percent of its initial value. L70 is the figure most often quoted. | Not a failure point. A fixture at L70 still emits light; it emits 30 percent less than it did when new. |
| ANSI/IES LM-80-15 | The approved method for measuring luminous flux and colour maintenance of LED packages, arrays and modules — the test that produces the raw data. | Not a lifetime rating. It reports measured maintenance over the test period, not a projection beyond it. |
| IES TM-21-11 | The approved method for projecting long-term lumen maintenance of LED light sources — the calculation that turns LM-80 data into a projection. | Not a statement about the luminaire. The IES states that the method applies only to the LED components within a luminaire. |
| The 6X rule | TM-21 projections are limited to 6 times the duration of the collected LM-80 data. | Not optional. The IES does not support marketing claims that exceed the 6X extrapolation limit and calls such claims misleading. |
The 6X rule: why a 100,000 hour claim is unsupported
The most consequential rule in this area is the limit the IES Testing Procedures Committee placed on its own model. The committee recognised that its models were limited in their ability to predict future performance, and therefore limited projections to six times the duration of the collected LM-80 data, in order to keep projections inside the model's statistical confidence levels.[1]
The example the IES gives is unambiguous: when TM-21 results are used correctly, projections using 10,000 hours of LM-80 data shall never exceed 60,000 hours, and in that situation a marketing claim of 100,000 hours to L70 would not be supported by IES standards. The position statement adds that TM-21 Addendum B, which replaced Addendum A, explicitly disallows reporting such projections, and that the IES does not support the use of marketing claims exceeding the 6X extrapolation limit.[1]
| LM-80 test duration collected | Maximum TM-21 projection permitted | Consequence for a datasheet claim |
|---|---|---|
| 6,000 hours | 36,000 hours | A figure above 36,000 hours is outside the permitted projection range. |
| 10,000 hours | 60,000 hours | The case the IES uses to reject a 100,000-hour L70 claim. |
| 12,000 hours | 72,000 hours | Longer test data is the only route to a higher supported projection. |
What horticultural luminaires actually report
The rule above describes what may be claimed. A separate question is what the horticultural lighting market actually reports, and there is a published survey that answers it. A 2020 review in the journal Sustainability compiled a database of 301 manufacturers and 1,473 LED lighting systems for plant growth, and reported the depreciation behaviour of the luminaires in it.[3]
The headline finding is that depreciation is measured in tens of thousands of hours, not in a single service life: the review states that the current LED luminaires' lifespans show 10 percent and 30 percent losses of light output after 45,000 and 60,000 working hours on average, respectively.[3]
| Operating hours | Average reported loss of light output | Interpretation |
|---|---|---|
| 45,000 hours | 10 percent | Roughly equivalent to L90 behaviour at that point. |
| 60,000 hours | 30 percent | Roughly equivalent to L70 behaviour at that point. |
Three further figures from the same review put the numbers in context. Lifespan data availability was assessed across 161 worldwide-distributed manufacturers, and L-standard categories were examined among the 116 companies that reported hours-rated life, covering 519 luminaires — so published rated life is a property of a subset of the market rather than of every product on it. The review also identifies Europe, at 40 percent, and North America, at 29 percent, as the main production regions for horticultural LED systems, and finds that the vast majority of the systems surveyed, about 70 percent, have efficacy values between 2 and 3 µmol/J.[3]
Turning rated hours into calendar years
Rated life is quoted in hours, but capital planning happens in years, and the conversion depends entirely on the photoperiod. A grow room running 12 hours a day accumulates a very different total from one running 20 hours a day.
| Photoperiod | Annual operating hours | Calendar time to 60,000 hours |
|---|---|---|
| 12 hours/day | 4,380 hours | about 13.7 years |
| 16 hours/day | 5,840 hours | about 10.3 years |
| 18 hours/day | 6,570 hours | about 9.1 years |
| 20 hours/day | 7,300 hours | about 8.2 years |
The table also shows why a long-hours operation should read a rated-life figure with more caution than a short-hours one. A fixture running 20 hours a day reaches 60,000 hours in roughly eight years, so the projection matters within the life of the equipment; the same fixture in a 12-hour greenhouse takes nearly fourteen years to get there.
What the standards do and do not cover
Two limitations are worth stating explicitly, because both are common sources of over-claiming.
The projection covers the LED, not the luminaire
The IES position statement is direct on this point: it is inaccurate to make marketing claims regarding the lifetime of an entire luminaire using only TM-21 projections, because that method applies only to the LED components within. The statement notes that a complete luminaire's reliability depends on many failure mechanisms beyond the LED component level, including drive circuitry and any secondary optical components, and that other testing standards focus on these degradation mechanisms and should also be used for determining complete luminaire lifetime. It concludes that the IES does not support the use of LED lumen maintenance life as the sole metric for determining solid-state luminaire lifetimes.[1]
Listing programmes add their own testing and reporting requirements
Third-party qualification programmes layer their own requirements on top of the base methods. Under the DesignLights Consortium horticultural programme, applications require that product testing be conducted at an accredited laboratory appropriate for the performance being evaluated, and the tests involved may include in situ measurement tests (ISTMT) and the ANSI/IES LM-79-19, LM-80 and LM-84 methods. Products must be complete LED light fixtures or modules to be eligible, and must hold safety certification from an organisation relevant in the United States or Canada.[2]
The programme's own timeline shows how quickly the qualifying population moves: Hort V4.0 applications started on 18 April 2025 with delisting on 5 January 2026, and the DLC reports that the average efficacy of listed products has risen 24.9 percent since Hort V1.0, with the qualified products list growing at over 102 percent a year since 2019. As of February 2025 the list comprised more than 1,200 V3.0 listed products from over 130 manufacturers, estimated to represent around 90 percent of the non-residential horticultural lighting market.[2] A product that was listed under one revision is therefore not automatically listed under the next one.
Planning replacement around PPFD, not hours
Because the practical effect of depreciation is a loss of light at the canopy, the replacement trigger should be expressed in the same units the crop plan uses. Four steps cover it.
1. Record the design PPFD and the daily light integral it produced
The installation was sized against a target. Whatever that target was, it is the reference value; a fixture decision made without it has nothing to be measured against.
2. Measure PPFD at canopy height, at several points
Uniformity matters as much as the average, so a single centre reading is not enough. Measuring at canopy height rather than at the fixture also captures the effect of any change in mounting height or canopy geometry.
3. Convert the loss into daily light integral
At a fixed photoperiod, a percentage loss in PPFD is the same percentage loss in daily light integral.
A measured 10 percent PPFD drop therefore removes 10 percent of the daily light integral the crop was receiving. If the original target sat near the lower end of the crop's recommended band, that loss can move the installation out of band well before any rated-life figure is reached; if the target sat in the middle of the band, there is more headroom.
4. Budget the driver and optics separately
Since the IES position statement identifies drive circuitry and secondary optical components as failure mechanisms outside the scope of TM-21, a replacement plan built only on the LED projection is incomplete. The two questions to ask a supplier are the LM-80 test duration behind the rated-life claim and what testing coverage exists for the driver and optical assembly.[1]
Frequently asked questions
Do LED grow lights burn out?
Usually not in the way a lamp does. LEDs lose output gradually rather than failing abruptly, which is why rated life is expressed as the projected operating hours until output falls to a stated percentage of its initial value. In a 2020 industry review of 301 manufacturers and 1,473 LED lighting systems, the compiled luminaire lifespans showed 10 percent and 30 percent losses of light output after 45,000 and 60,000 working hours on average respectively.[3]
Is a 100,000 hour L70 rating credible?
Not when it is derived from a short LM-80 test. The Illuminating Engineering Society limits TM-21 projections to 6 times the duration of the collected LM-80 data, so projections from 10,000 hours of LM-80 data must never exceed 60,000 hours. The IES states that a marketing claim of 100,000 hours to L70 in that situation would not be supported by IES standards, and that TM-21 Addendum B explicitly disallows reporting such projections.[1]
Does TM-21 predict the life of the whole fixture?
No. The IES states that it is inaccurate to make marketing claims about the lifetime of an entire luminaire using only TM-21 projections, because the method applies only to the LED components inside it. Luminaire reliability also depends on other failure mechanisms including drive circuitry and secondary optical components, which are covered by different test standards.[1]
How do I know when to replace a grow light?
Measure rather than count hours. Because a fixture's depreciation shows up directly as a loss of photosynthetic photon flux density at the canopy, the decision point is the moment the measured PPFD at canopy height drops below the value the crop plan depends on. A 10 percent fall in PPFD at a fixed photoperiod is a 10 percent fall in daily light integral, so the replacement threshold should be tied to the daily light integral target that was used to size the installation.
About the publisher
Zhongshan City Ruixian Electronics Factory (XineLam) designs and manufactures LED lighting products and LED grow light modules. The company has 17 years of experience in the LED lighting industry and holds 300+ patents in China and internationally.
Rated-life data and thermal design are decisions made at the module and fixture level, and both are part of how we specify a product conversation. Where a project is planned around a rated-life target, the figures worth requesting are the LM-80 test duration behind the claim, the projected maintenance curve rather than a single number, and the separate coverage for the driver and optical assembly described above.
This page is a technical explainer assembled from the standards and peer-reviewed sources listed below. The depreciation values quoted are published averages and programme thresholds, not measurements of any specific product, and nothing here is a performance claim for a particular fixture.
References
- Illuminating Engineering Society. PS-10-18: IES Position on LED Product Lifetime Prediction, issued 9 October 2018. https://ies.org/advocacy/ps-10-18/ (accessed 11 September 2026).
- DesignLights Consortium. Horticultural Technical Requirements V4.0, effective 18 April 2025. https://designlights.org/our-work/horticultural-lighting/technical-requirements/hort-v4-0/ (accessed 11 September 2026).
- Paucek, I., Appolloni, E., Pennisi, G., Quaini, S., Gianquinto, G. and Orsini, F. LED Lighting Systems for Horticulture: Business Growth and Global Distribution. Sustainability 12(18): 7516, 2020. https://www.mdpi.com/2071-1050/12/18/7516 (accessed 11 September 2026).
Note on scope: this guide describes how rated life is defined, projected and limited under the cited standards, and how to turn those figures into a measurement-based replacement plan. It does not recommend a specific product, and the depreciation percentages quoted are published industry averages rather than a guarantee for any fixture.