The Obsolescence Clock on Your Installed Base

What DOE Actually Did in February 2026

Start here, because this is the fact that separates a credible obsolescence brief from a content farm.

On February 13, 2026, the Department of Energy published a final determination on energy conservation standards for metal halide lamp fixtures (91 FR 6737, effective March 16, 2026). DOE concluded that more-stringent standards would not be cost effective, and therefore declined to amend the existing rules.

The current standards — minimum ballast efficiency levels at 10 CFR 431.326, applying to fixtures designed to operate 50W to 1000W lamps, in force for units manufactured on or after February 10, 2017 — stay exactly as they are. This was the three-year lookback following a similar no-amendment determination in 2021.

So: no federal tightening, no federal ban, no 2027 cliff. Anyone telling you otherwise has not read the docket.

Now the part that makes this interesting rather than merely deflating. Buried in the same rulemaking record is DOE’s own shipment modeling, which projects a decline of more than 99% in metal halide lamp fixture shipments by 2030 relative to 2021.

Read those two facts together and the strategic picture is clear. The federal government declined to regulate a product category because it expects that category to disappear on its own. Regulatory inaction is not a reprieve. In this case it’s a forecast — and the forecast is that the market evaporates regardless.

That is the correct frame for your installed base. Not “when does this become illegal,” but “when does the last manufacturer stop running the line.”

Three Separate Clocks, Running at Different Speeds

The obsolescence pressure comes from three independent mechanisms. They get conflated constantly, and conflating them is exactly how you end up with a wrong number in a board deck.

Clock One: State Sales Bans on Fluorescent

This one is real, active, and specific. States including California, Vermont, Colorado, Oregon, Washington, Rhode Island, Hawaii, Minnesota, and Maine have enacted laws restricting or banning fluorescent lamp sales, with Illinois joining as the tenth. Timelines run roughly from 2024 through 2029 depending on the state and lamp type, and they typically arrive in tiers — screw- and bayonet-base CFLs first, then pin-base CFLs and linear fluorescent tubes a year or so later.

Two things about these laws that are consistently misreported:

• They restrict sales, not use. No state is requiring you to remove installed fixtures. You may keep running what you own until it fails. What you lose is the ability to buy a legal replacement lamp inside that state.
• They are fluorescent laws. They do not ban metal halide, high pressure sodium, or mercury vapor. Anyone extending the fluorescent ban list to HID is fabricating.

It’s also worth understanding the tiering, because it explains the shape of the supply curve. Most of these statutes hit screw- and bayonet-base CFLs first, pin-base CFLs and linear tubes twelve to twenty-four months later. That staggering was designed to give the market time to adjust. In practice it gave manufacturers a clear, published schedule for winding down production — which they have followed.

Clock Two: RoHS Exemption Expiries

The EU’s RoHS Directive prohibits mercury in electrical equipment unless a specific exemption applies. Lamps run on exemptions in Annex III, each with a stated expiry:

And here is the nuance that almost nobody gets right. A listed expiry date is not a phase-out date. As LightingEurope has stated plainly, exemptions can be renewed; if a renewal application is filed, the exemption remains valid past its listed expiry while the Commission evaluates it; and if renewal is ultimately refused, the Commission must still grant a transition period of roughly 12 to 18 months.

So “metal halide is banned in February 2027” is not a defensible claim. What is defensible: the exemptions are under sustained pressure, each renewal cycle narrows the scope, and every narrowing removes another slice of European volume from a global manufacturing base.

That last clause is the point. These are EU rules. They don’t govern a Texas refinery. They govern whether the factory that supplies that refinery still exists.

Clock Three: The Minamata Convention

The international mercury treaty is the biggest lever and the most frequently misstated. At COP-4 in 2022, parties agreed to phase out integrated-ballast CFLs after 2025. At COP-5 in November 2023, they amended Annex A to phase out all fluorescent lamps for general lighting purposes by 2027.

Note the scope carefully: fluorescent lamps for general lighting. Metal halide and high pressure sodium are not on that phase-out list. If you have seen a chart showing Minamata killing HID in 2027, that chart is wrong.

What Minamata does to your HID supply is indirect but powerful — it removes the fluorescent volume that shared factories, glass capacity, mercury-dosing equipment, and distribution shelf space with HID product. Categories don’t die alone.

Why Nine States Create a Fifty-State Shortage

This is the mechanism that matters most to a maintenance manager in Oklahoma or Louisiana, and it has nothing to do with the law in their state.

Lamp manufacturing is a volume business built on long production runs, dedicated tooling, and thin margins. When a manufacturer loses California — plus Illinois, Washington, Colorado, and the rest — it doesn’t lose a slice of demand. It loses the volume threshold that made a SKU worth producing at all.

Think about it from the plant manager’s side. A production line that ran three shifts to serve fifty states now serves forty. Tooling amortization, mercury-dosing equipment, quality systems, and regulatory documentation are all largely fixed costs; they don’t shrink by twenty percent when the market does. The unit economics invert well before the demand does.

The rational response is not to keep the line running for the remaining states. It’s to consolidate SKUs, then discontinue the category. The result is that the sales ban in Sacramento shows up as a twelve-week lead time in Baton Rouge, and eventually as a discontinuation notice.

You will see it in this order:

1. Lead times stretch. What was on the shelf becomes a special order.
2. SKU rationalization. Wattages, color temperatures, and base types quietly disappear. The specific lamp your fixture was certified with is often among the first to go, because specialty variants have the thinnest volume.
3. Price decoupling. Remaining stock prices rise against no underlying cost change. Grey-market and new-old-stock sellers appear.
4. Not announced with a deadline. Announced by a distributor telling you they can’t get it.
5. Ballast scarcity, which is worse. Everyone plans for lamps. Almost nobody stocks ballasts. A ballast failure on a discontinued platform is where a lighting problem becomes a fixture replacement problem — and a fixture replacement in a live process unit is a multi-thousand-dollar event, not a parts swap.

Why Hazardous Location Is Exposed First

Every dynamic above hits classified-area lighting harder than general industrial lighting, for one structural reason: hazloc lamps were already low-volume specialty SKUs.

A 400W metal halide for a warehouse high bay is a commodity. The specific lamp qualified inside your explosionproof luminaire — right envelope, right base, right operating position, right ANSI ballast designation — is not. It exists as a variant of a variant. When rationalization comes, thin-volume specialty SKUs go first, and they go without notice.

There is a compliance dimension too, and it’s easy to miss. Your luminaire’s certification and its temperature classification were established with a specific lamp in it. When your usual lamp is unavailable and a distributor offers a substitute with the same wattage and base, that substitute is not automatically a legal replacement. Different arc tube geometry, different envelope, different thermal behavior — and your T-code was a test result for the original configuration, not a property of the housing.

The practical trap: a substitution made at 2 a.m. by a shift electrician trying to restore light to a unit is a documented deviation nobody will catch until an audit — and a lamp that isn’t the certified lamp is not a replacement in kind. Scarcity doesn’t just cost money. It generates compliance drift.

The Stockpile Option, Honestly Assessed

The first instinct of every experienced maintenance manager hearing this is reasonable: buy ten years of lamps now and stop worrying about it.

Sometimes that’s the right call. Often it isn’t, and it’s worth being straight about why.

What stockpiling solves. It buys time on a known SKU at a known price, and it converts an unpredictable availability risk into a predictable storage cost. For a small population of fixtures in a hard-to-reach location that you’re not ready to convert, a strategic buy is genuinely sensible.

What it doesn’t solve:

• A lamp stockpile protects against half the failure modes, and the power supply is the component that actually fails. Ballast electrolytics degrade on the shelf as well as in service, so a ten-year ballast stock is not a ten-year ballast supply.
• The access cost. Stockpiling locks in the recurring relamp cycle — and every relamp needs the same scaffold, the same permits, and the same isolation as a conversion would have. You’ve bought lamps, not labor.
• Capital and space. Ten years of lamps is dead capital plus warehouse volume, and mercury-containing lamps carry universal waste handling obligations under 40 CFR Part 273 whether they’re installed, stored, or eventually discarded.
• Degradation you’re already paying for. HID lumen output falls substantially over life, so a stockpile commits you to another decade of declining illuminance in areas where light level is a safety control.

The honest version: stockpile deliberately, for a defined and inventoried subset, as a bridge to a conversion you’ve already scheduled. Stockpiling as an alternative to deciding is how sites end up with eight hundred lamps for fixtures whose ballasts are unobtainable.

How to Actually Assess Your Exposure

Obsolescence risk is measurable. Most sites have never measured it.

Build the inventory. Fixture population by area classification, lamp type, wattage, ballast platform, and — critically — by specific certified lamp SKU. Not “metal halide.” The actual part number on the nameplate.

Score each SKU on three axes. Current availability and lead time; whether the manufacturer has issued rationalization or discontinuation notices; and whether the SKU has any second source. A part with one manufacturer and a rising lead time is your top risk regardless of how many spares you hold.

Count the ballasts separately. Lamp availability and ballast availability are different curves and ballast is the steeper one.

Map the SKU risk against access difficulty. A fixture that’s easy to reach and hard to source is a stocking problem. Substituting a different lamp to get out of trouble is not a free move either, since the certification was established with a specific lamp in the housing. A fixture that’s hard to reach and hard to source is a capital planning problem, because when it fails you’ll be buying access, permits, and a crew — not a lamp.

Then plan against turnaround windows, not fiscal years. If replacement work needs hot work and therefore a turnaround, your real decision horizon is the next window, not next quarter. Miss it and you’ve added a full cycle.

Reframing the Buying Trigger

Energy payback has carried the LED conversation for fifteen years, and in hazardous areas it has always been the weakest argument — the energy delta on a single fixture rarely justifies the installed cost on its own.

Supply risk is the stronger case, and it’s the honest one:

The question is not whether LED pays back on kilowatt-hours. It’s whether you want to discover your lamp is discontinued during an unplanned outage, or on a schedule you chose.

That framing survives scrutiny in a way “2027 HID BAN” does not. It doesn’t require you to overstate any regulation. It rests on a documented federal shipment forecast of a 99%-plus decline, ten states removing fluorescent volume from a shared manufacturing base, EU exemptions narrowing at each renewal, and a treaty phase-out that takes the adjacent category off the same production lines.

None of that makes your fixtures illegal. All of it makes them orphans.

The organizations that handle this well are not the ones that panicked about a ban. They’re the ones that inventoried their SKUs, found the twelve part numbers with a single source and a stretching lead time, and put those fixtures in the next turnaround scope while they still had a choice about the timing. If conversion is the answer, the listing question is the first one to settle.

Sources and further reading: DOE final determination, 91 FR 6737 · DOE metal halide lamp fixtures program page · ASAP standards summary · Minamata Convention amendments · EU RoHS Directive · Commission Delegated Directive (EU) 2022/283 · State fluorescent legislation overview · EPA universal waste lamps