I handle component procurement for a 140-person lighting manufacturer. In 2024 I signed off on roughly $310,000 in LED drivers, power supplies, and control gear across four product lines and eleven suppliers. I've been doing this since 2019.
The cheapest driver on your quote is almost always the most expensive part of your finished fixture.
That's not something I picked up from a vendor deck. It's the conclusion I got to after watching the same mistake get made three times — twice by my predecessor, once by me.
One caveat before I go further: I'm not a power electronics engineer. I can't explain why one driver topology shrugs off an inrush event while another one welds its relay shut. What I can tell you is what shows up on our side of the invoice when it does.
The Number You Compare Isn't the Cost You Pay
In 2023 we bid out a 40W constant current driver for one of our outdoor lines — seven vendors, same nominal specs, quotes between $8.40 and $13.10 per unit (this was back in 2023, so the pricing's moved since). I went with an $8.40 unit. My reasoning was defensible: same current output, same forward voltage window on paper, same UL listing claim. At 4,000 units a year, that's a $7,200 spread against the mid-tier option at $10.20.
What the quote didn't include was the second shipment. It never does.
Failures started showing up in month seven. We ran around 2.6% by eighteen months — I'd have to pull the exact figure, but it was just under 3%. Out of 4,000 units that's roughly 104 drivers. Not all of them turned into a truck roll, but enough did. Commercial outdoor failures mean an electrician, sometimes a lift, sometimes a second trip. Our average site visit ran about $325 (drive time, minimum labor, trip charge — that's before the driver itself). We logged roughly 60 of those, plus another 40-odd units handled as straight RMA with freight and admin, call it $45 each. None of that ever appears on a purchase order.
Rough cleanup total: $22,000. Against $7,200 in savings.
Most buyers focus on the per-unit price and completely miss the cost of the second shipment. Then they miss the third thing, which is worse.
When the Driver Dies, Nobody Blames the Driver Manufacturer
In Q3 2024 we retrofitted a 60-fixture parking garage. Two drivers failed inside the first month (ugh). Same cheap batch, different product line. The building manager's takeaway wasn't "that driver brand is unreliable." It was "your fixtures are unreliable." He was standing in the dark looking at our label.
That's the part that doesn't fit in a spreadsheet. When you buy a constant current driver for resale under your own name — which is exactly what a private label or LED power supply OEM program is — you aren't buying a component. You're buying the failure rate your customers will attribute to you. The contractor on that garage job didn't call us for the next bid. I can't put a clean dollar figure on that, and I've stopped trying.
So I don't really buy drivers anymore. I buy the number of times my phone rings about a fixture.
And the causation runs in the direction people get backwards. Vendors who can prove quality charge more, so their price looks like the cause of the quality. It isn't. The price is the signal. The quality is the reason they can charge it.
What to Actually Compare: A Working Constant Current Driver Specification Guide
Since the difference lives in the spec sheet, here's what I ask for now. Nothing exotic — just the things that separate an $8.40 driver from a $10.20 driver when the labels look identical.
- Output current tolerance and ripple. ±5% with no ripple spec is a different product from ±3% with ripple under 5%. Current ripple turns into lumen depreciation and color shift later, when it's no longer the driver supplier's problem.
- Forward voltage window across temperature. A narrow Vf range that only covers your array at 25°C is a warranty claim waiting for a cold morning.
- Inrush current. NEMA 410 exists because high-inrush drivers cook occupancy sensors and relays on the same circuit. Ask for the number, not the adjective.
- Surge immunity rating. The kV figure and the test waveform (IEEE C62.41.2 combination wave vs. ring wave) matter far more than the phrase "surge protected."
- Isolation and SELV, IP rating, and the TC point on the case — with the derating curve that goes with it. If there's no derating curve, there's a fixed ambient assumption buried in the fine print.
- Dimming compatibility. 0-10V means different things on different drivers, and the only test that counts is your control on your fixture. Universal compatibility isn't a real thing — verify it on the bench before you commit to a PO.
- Certification, and who holds the file. UL 8750 for the US, EN 61347-1 and EN 61347-2-13 for the EU, IEC 61347-2-13 internationally. In the EU, Regulation 2019/2020 has applied to separate control gear since September 1, 2021 — check the no-load and standby limits against the current text at the official source, since it's been amended.
- Warranty exclusions. The useful question isn't how many years. It's what voids it.
The question everyone asks is "what's your best price?" The question they should ask is "what voids the warranty, and who pays the freight on a replacement?"
The Counterargument, Because It's Fair
There are plenty of places where the budget driver is the right call. A 12W indoor downlight on a spec house with a paint-to-sale cycle. The replacement market, where the buyer knows the exact model number and just wants it cheap. Low-current decorative work in conditioned space. I've bought the cheap one on purpose more than once. The rule isn't "always buy up." The rule is: know which one you bought, know why, and price the risk in.
I also want to be honest about the data, because this is where most procurement arguments fall apart. I don't have industry-wide failure numbers — nobody publishes clean ones, and the figures vendors quote are marketing. What I have is our own records across about 14,000 drivers bought between 2021 and 2024, and in that sample the cheapest pricing band ran roughly four times the first-year claim rate of the mid band. One company, one product mix. Treat it as directional, not gospel.
And it's worth limiting the scope: all of this comes from a component buyer's seat at a mid-size OEM. If you're moving 200,000 units a year, your calculus on vertical integration and incoming inspection looks nothing like ours.
What Changed in Our Process
Our procurement policy now requires quotes from three vendors minimum, and no driver gets approved off a datasheet. We burn in 20 samples at rated load for 72 hours before a first order. For private label programs we inspect the physical label and confirm the model number matches the certification file, because a label change that isn't covered by the file is a compliance problem waiting to ship.
Moso has stayed on our shortlist since 2022 for the outdoor line. They aren't the cheapest quote we get — and I've stopped expecting that to change, because I've never once picked a supplier on price alone and kept them. What kept them there is that the moso led driver units we've run came in under 1% on first-year claims, and their OEM team could put our label on the same listed platform without dragging us through a re-certification cycle. That last part is worth more to us than the $0.90 a unit we'd save elsewhere. I want to say we've bought around 11,000 of them over three years, but don't quote me — it's closer to 9,000 once I exclude the 2023 order we cancelled.
The moso driver platform isn't magic. It's a driver that does what its spec sheet says, backed by a manufacturer that answers questions in writing. That's a low bar that a surprising number of suppliers only clear after you've paid them.
If you're an LED driver distributor deciding what to stock, or a lighting OEM putting your name on the label, remember that the driver is the cheapest brand insurance you'll ever buy — and the most expensive thing you can cut. Price it on total cost. The unit price was never the interesting number.
