Are Most Horticultural Lighting Systems Optimised for Plants or for Specification Sheets?
Walk into any lighting procurement conversation in CEA and the comparison usually comes down to a handful of numbers: efficacy, power consumption, output, channel count, and above all, cost. These figures are clean, comparable and easy to put in a table. They’re also, on their own, a poor predictor of what actually happens to a crop once that fixture is installed.
That gap is worth sitting with, because it points to something structural in how this industry designs and sells light.
The metric that wins the tender isn’t the one that wins the harvest
Efficacy is fundamentally an electrical engineering measure: how many photons come out per watt in. It’s a legitimate number, and it matters for running costs. But it says nothing about the quantity or quality of light reaching the canopy level, how photons are distributed across the growing areas at real mounting heights and installed spacings, how output and spectrum drift as a fixture ages, or how a given spectrum interacts with a specific crop’s morphology across germination, vegetative growth, flowering and fruiting. Photobiology doesn’t respond to a lab-rated number. It responds to what actually lands on the leaf, consistently, over months, and that response is entangled with every other environmental limiting factor in the room. Light isn’t a purely additive input: applied wrong, it’s just as capable of inducing stress and stunting growth as it is of encouraging it.
The reason efficacy still dominates the conversation isn’t that it’s the most important variable. It’s that it’s the easiest one to measure, print on a datasheet and win a comparison with. Manufacturers optimise for what can be benchmarked and marketed, and a single headline figure benchmarks and markets very well. Plant productivity, distribution uniformity, and real-world consistency don’t reduce to a single number, so they lose the comparison even when they matter more.
Plants don’t lie
Whatever a spec sheet says, the crop is the final and only honest arbiter. You can sell capability, but you can’t sell yield, morphology, or consistency, those either show up on-site or they don’t. And what determines whether they show up is rarely the headline number. It’s the accumulation of details that don’t make it onto a datasheet. Uniformity of PPFD across the canopy rather than at a single test point, how spectrum shifts across the dimming range, whether photoperiod and DLI stay consistent through the full crop cycle each season, how the fixture behaves at the edges of a bay rather than dead centre, how serviceable and reliable it is years later rather than on day one.
None of these individually looks like a big deal on a spec comparison. Cumulatively, they’re the difference between a system that performs on paper and one that performs in a bay full of live plants.
A pattern inherited, not invented
This isn’t a new failure mode for lighting. It’s an old one wearing a new unit of measurement. A large share of horticultural lighting manufacturers, and a large share of the engineers designing for them, came out of general lighting: street lighting, retail, commercial interiors. I did too, at Osram and GE, before moving into CEA. And the longer you spend in that world, the more familiar this pattern looks.
Street lighting spent decades optimised around lm/W and initial lumen output, because those were the numbers that won tenders and satisfied a photometric compliance sheet. I watched schemes get specified and approved entirely on paper, on those numbers, and then get installed and immediately generate complaints: glare, poor uniformity between poles, colour rendering that made a road look fine in a datasheet and lifeless at night. The compliance number was satisfied. The lit environment wasn’t. Nobody was measured on whether the road was actually easier to drive at night, so that’s not what got engineered for.
CEA lighting is running the same playbook with a different unit. Swap lumens for micromoles and underneath both sits the same real target, the lowest cost per photon delivered, and the mechanism is identical: a headline figure that’s cheap to measure and easy to compare wins the spec, while the qualities that determine real-world performance, uniformity, consistency, spectrum, behaviour at the edges of a bay, don’t show up until the system is running. The industry didn’t invent metric-chasing when it moved into horticulture. It imported it, largely because the people and companies building the fixtures were the same people and companies who’d spent careers building streetlights, office panels and downlights to the same incentive structure.
The difference is what’s on the other end of the failure. A poorly uniform car park is an inconvenience. A poorly uniform canopy is a yield variance you don’t see until harvest.
A symptom of designing with blinkers on
A lot of this comes down to a narrower failure: treating lighting as an electrical efficiency problem rather than a horticultural systems problem. Efficacy-first design optimises one variable in isolation and treats spectral engineering, thermal behaviour, mechanical layout and grower usability as secondary concerns to be resolved after the electrical spec is locked in. It also sidelines advanced control, and with it plant light-use efficiency, in favour of chasing fixture light-output efficiency instead.
That ordering shows up in the field in a familiar pattern. Facilities install the newest dynamic, multi-channel fixtures, marketed on their ability to tune spectrum precisely for every growth stage, and end up running every channel at 100%, functionally identical to the static fixture it replaced. That’s rarely carelessness on the grower’s part. It’s usually the result of a system designed around what a fixture could technically do, with too little thought given to whether the people running the facility had the time, training or interface to actually use it. Capability that isn’t usable in practice isn’t capability delivered; it’s a number that helped win the sale.
What a holistic approach actually asks for
Optimising for the plant instead of the spec sheet means starting the design from crop response and operational reality, not from the fixture’s electrical performance. It means treating light, electrical, mechanical and control systems as one integrated design problem rather than a sequence of separate disciplines handed off to each other. It means validating claims against real distribution and real crop outcomes, not just lab-rated output. And it means being honest that a feature only has value if it survives contact with a working facility and the people who run it day to day.
The people on the other end of the spec sheet
Some of the most valuable time I’ve spent in this industry hasn’t been in a lab or a design review, it’s been on-site in conversation with growers and researchers who’ve spent decades learning what a crop actually needs, and who are remarkably generous about passing that knowledge on. This industry has more of that spirit than most: people who’ve made every mistake once, know exactly where the theory breaks down against reality, and would rather hand that knowledge to the next generation than guard it. Every genuinely good lighting decision I’ve been part of has come from listening to those people, not from a datasheet.
That matters because of who’s on the other end of the installation. A lot of the growers I work with are running family businesses, farms where the operation’s performance isn’t an abstract KPI, it’s the household income, the next generation’s inheritance, a livelihood tied directly to whether the crop comes in on target. When a fixture underperforms its spec in the field, that’s not a line item for a facilities budget somewhere. It’s a real cost carried by real people. That’s a different level of accountability than a lighting spec written for a car park or an office ceiling, and it’s one the industry should hold itself to more visibly than it currently does.
Where that leaves the question
I don’t think the industry is choosing spec sheets over plants out of bad faith. I think it’s optimising for what’s easy to measure, easy to compare, and easy to sell, and the plant’s actual response is none of those things until months into a crop cycle. Closing that gap means manufacturers being more honest about the difference between rated and delivered performance, and it means designing holistically, where technology meets biology, rather than defending a headline efficacy number and hoping the rest works itself out.
The plants will tell you either way. The only question is whether the system was designed to listen.
