US Flashlight Market Report, Size, Share and Forecast 2026-2033: Tariffs, Rechargeables and Amazon

Three forces, one line item

Tariffs, rechargeable cells and Amazon are usually told as three separate stories about the same product. In the United States they converge on a single line item - the gross margin between what a light costs landed and what it sells for - and each of them compresses it. None of them expands the category. That observation drives every number in this report, and it is why a market exposed to three large forces still grows at 3.7 per cent.

The tariff story is not a rate going up. It is a rate that went up, was struck down by the Supreme Court on 20 February 2026, was replaced within four days under a different statute, and settled above where it began - while the duty on flashlight parts stayed more than twice the duty on a finished flashlight. The rechargeable story is not batteries getting better; it is a product that used to generate a repeat purchase and now does not. The Amazon story is not e-commerce share; it is a category where the buyer searches for a lumen figure rather than a brand, which turns the shelf into a price list.

What this report counts, and what it leaves out

This is a national report for the United States. It counts portable lighting sold as a product: handheld flashlights, headlamps, lanterns, spotlights and work lights, right-angle and pen lights, and the safety-rated lights sold into fire, utility and industrial channels. It counts them at first-sale value - the price at which the importer or the domestic manufacturer disposes of the unit, before retail margin.

It deliberately excludes three adjacent things. Vehicle lighting is a different tariff heading (8512) and a different buyer. Fixed and architectural lighting is heading 9405 and does not travel. And a light shipped inside a cordless-tool kit is counted in the tool, not here - a real omission, quantified in the technology chapter, and the single largest blind spot in this perimeter.

Two of those three lines are official. The domestic shipment figure is not. No agency publishes flashlight production separately, so it is built from the named US assemblers and their published product ranges, and a reader who disagrees with it should move the whole forecast by the size of the disagreement rather than by a percentage of it. The retail line appears because this market is normally quoted at retail, and it is labelled because the multiplier is an assumption, not a measurement.

The only hard number in this market is the import bill

In 2025 the United States imported USD 770 million of portable lighting and exported USD 140 million. China supplied USD 618 million of those imports, 80.26 per cent of the total. Vietnam supplied 78.8 million, Thailand 23.7, Cambodia 14.5 and Mexico 7.2. The three Asian origins after China total 117.0 million, or 15.19 per cent, and Vietnam and Thailand were the two fastest-growing origins of the year - the sourcing shift that every tariff analysis of this category predicts is already visible in the 2025 data, one year before the current duty schedule took effect.

Run the monthly series forward and the shift sharpens rather than fades. US imports of the heading were USD 76.1 million in August 2026, which is 1.186 times the 2025 monthly average - 18.6 per cent above the run rate of a year in which duties were lower. A single month is not a trend, and August carries hurricane-season and back-to-school demand, but it is enough to falsify the simplest prediction about tariffs, which is that they reduce the volume of taxable imports. At a 25 per cent rate the duty on the Chinese share alone was about USD 154.5 million on 2025 volumes.

The tariff stack, and the inversion inside it

The rate that matters is not the one on the finished product. It is the ratio between the rate on the finished product and the rate on its inputs, because that ratio decides where the value is added.

Three numbers follow, and they are the report's spine. One caveat on the parts line, stated because it bounds the arithmetic: the metals-derivative component is commonly assessed on the aluminium or steel content value rather than on the whole customs value, so a part in which metal is a minority of its cost pays less than the full 62.5 per cent. The comparison that matters survives that caveat, because the finished line carries no metals-derivative duty at all - the parts route is disadvantaged by at least 25 percentage points, and by more than that wherever metal is a material share of value. The 231.75 million figure below is therefore an upper bound on the classification gap, not a point estimate.

• Parts cost 2.50 times what the finished light costs in duty. A body, a bezel, a reflector, a tailcap switch and a lens assembly imported into the United States pay 62.5 per cent. The same components, assembled into a working flashlight in Shenzhen and shipped as one article, pay 25 per cent. The tariff schedule therefore charges a 37.5 percentage point premium for the act of assembling the product in America.
• A cell costs 2.14 times what the finished rechargeable light costs in duty. This is the second half of the same inversion, and it is newer: the Section 301 increase on non-EV lithium-ion cells to 25 per cent took effect on 1 January 2026, and the auto-parts action stacks on top of it.
• The wedge is worth USD 231.75 million a year. That is 37.5 percentage points applied to 2025's Chinese import bill, and it is equal to 26.9 per cent of the entire first-sale market. It is not a tax on the category; it is a number roughly the size of the category's whole domestic manufacturing base, decided by which ten-digit code a customs broker selects.

The practical reading is uncomfortable for anyone who assumes tariffs are a reshoring instrument. For this product the schedule subsidises the import of finished goods and taxes the import of the inputs that would be used to make them here. A US assembler importing parts pays 62.5 per cent; a competitor importing a boxed light with a lithium cell inside pays 25 per cent on the whole thing. The US firms that do assemble here describe the model plainly - SureFire states that its illumination products are "designed, perfected, and manufactured in the USA with the finest materials and components sourced worldwide" - which is exactly the position the schedule penalises.

How the tariff survived its own repeal

The 25 per cent is not the high-water mark, and the route to it matters more than the level. Chinese flashlights entered at 12.5 per cent before 2025. From February 2025 the IEEPA fentanyl duty and, from April, the reciprocal tariff took the stacked rate as high as 42.5 per cent on the heading, easing to 32.5 per cent when the fentanyl component was halved in November 2025. On 20 February 2026 the Supreme Court held in Learning Resources v. Trump that IEEPA does not authorise tariffs at all, and both sets of duties fell away - with a refund process opened for importers who had paid them.

Four days later the administration imposed a 10 per cent surcharge under Section 122 of the Trade Act of 1974, a bridge statute limited to 150 days. Then, on 24 July 2026, the Section 301 forced-labour action took effect across 60 economies at 10 or 12.5 per cent, and Section 122 lapsed. The net result: 25 per cent, 7.5 points below the 2025 peak, on a completely different statutory footing, with no transition period a purchasing team could plan around and no guarantee of duration. Two of the three rates in the table above were established within the last ten months.

That is the actual tariff risk in this market - not the level, but the fact that the legal basis has changed three times in eighteen months and the refund channel is open on the first of them. A tariff whose collection can be reversed is a tariff a buyer discounts, and a discount that arrives late arrives as a windfall rather than a price cut.

What the three forces do to one unit

Take a rechargeable 21700 work light that lands at USD 12.00 and retails at USD 39.99, and the same light built under a US assembly model from imported parts.

• Imported finished: duty of 25.0 per cent, USD 3.00, and a landed cost of USD 15.00.
• Assembled domestically: the housing and optical parts alone carry 62.5 per cent, and the cell carries 53.4 per cent. Even on a conservative split - USD 6.00 of parts and USD 4.00 of cells - the duty is USD 3.75 and USD 2.14, or USD 5.89 against USD 3.00, with the labour of assembly still to be paid.

That is the whole argument in one comparison. The tariff does not make domestic assembly competitive; it makes it about USD 2.89 a unit worse before a single wage is paid. Which is why the domestic share of this market has not moved, and why the investment chapter finds no one building capacity.

Amazon then sets the revenue side against that cost base. E-commerce accounted for 17.1 per cent of US retail sales in the second quarter of 2026, growing 12.2 per cent year on year against 6.7 per cent for retail as a whole. In a category where a buyer types "2000 lumen rechargeable" rather than a brand, that channel does not merely distribute the product; it publishes the price of every alternative simultaneously, and the brand with the 25 per cent cost disadvantage is the one whose listing cannot hide it.

Forecast

The forecast is a build rather than a curve fit: volume, then price and mix, compounded from a 2025 base of USD 860.0 million to 2033.

The 3.7 per cent is a nominal rate and most of it is price, not appetite. At 3.0 per cent - the case in which refunds are passed through and lumens per dollar keep falling - the 2033 market is USD 1,089.4 million, and that 61 million gap is the honest error bar on this forecast. Volume is the number to attack, because the strongest structural force in the category - a rechargeable light that runs for a decade on a cell that costs less than a pack of alkaline batteries - is a force that lengthens replacement cycles.

What would change this view

• The parts rate falling below the finished rate. A single line in a future Section 301 or Section 232 action could invert the inversion. If flashlight parts were removed from List 3, domestic assembly stops being a 62.5 per cent business and the import bill stops being the whole story.
• A channel that publishes unit volume. Nothing in this perimeter is measured in units by any citable source. The moment one retailer or one trade body does, the 1.0 per cent volume assumption becomes checkable and the price-mix split stops carrying the forecast.
• Tool-platform capture. If cordless-tool makers keep absorbing work lighting into battery ecosystems, this market's addressable perimeter shrinks by a channel rather than by a share - and the report would be measuring the residue.

Until at least the first of those happens, the defensible statement about US portable lighting is not that it is growing. It is that it is being re-priced, by three forces that all take their cut at the same place.

Table des matières

1. Market Overview
   • Three forces, one line item
   • What this report counts, and what it leaves out
   • The only hard number in this market is the import bill
   • The tariff stack, and the inversion inside it
   • How the tariff survived its own repeal
   • What the three forces do to one unit
   • Forecast
   • What would change this view
2. Competitive Landscape
3. Value Chain
4. Investment Activity
5. Technology and Innovation
6. Market Risk
7. Regulatory Landscape

Paysage concurrentiel

Two markets wearing one heading

The US flashlight market is not one market. It is a professional market with a spec and a procurement officer, and a household market with a drawer and a price tag, and they share a tariff code and nothing else.

The professional market is where the price is. Weapon-mounted lights, helmet lights, right-angle fire-service lights, ATEX-rated confined-space lights and scene lighting are bought by agencies against standards, replaced on schedules, and rarely shopped on price. The brands that hold it are American and they say so. SureFire states that its illumination products are designed, perfected and manufactured in the USA with components sourced worldwide. Streamlight organises its catalogue around safety ratings and standards – ATEX Zone 0 on the Dualie, ANSI information as a standing resource – which is what a spec-driven buyer needs and what a consumer never reads. Maglite sells the same heritage into the household, with outage and emergency as an explicit category rather than an accident of the range.

The household market is where the volume is. It is won on lumens per dollar, on a shelf position, and increasingly on a search result. It is served by imports, by Amazon-native brands that have no US distribution cost at all, and by the battery brands whose name on a blister pack is worth more than any specification inside it.

The seven credible multi-line suppliers

This report counts seven suppliers with a national presence across more than one product family: Energizer, Mag Instrument, Streamlight, SureFire, Coast, Pelican and Princeton Tec. That is the count of credible multi-line participants, not a share, and it is deliberately not a concentration measure – none of the seven is close to controlling the category, and the top of the market is more fragmented than the count suggests because the volume tier below them is hundreds of importers and marketplace sellers with no brand at all.

Three structural facts hold the landscape where it is:

• No one consolidates. There has been no transformative merger in US portable lighting in this cycle. The brands that could have bought the category did not, because the category’s margin is set by a tariff schedule and a marketplace rather than by scale.
• The battery brands are not really in it. Energizer’s lights sit inside a battery business, and the segment’s disclosure shows exactly what that means for pricing power. In fiscal 2025, ended 30 September, Energizer reported net sales of USD 2,952.7 million, up 2.3 per cent, with fourth-quarter organic sales down 2.2 per cent on volume down 2.9 per cent, partly offset by price up 0.7 per cent driven by innovation and tariffs. Price is doing the work; volume is not. A category where the price line is a tariff pass-through and the volume line is negative is not a category anyone is racing to buy.
• The tool platforms are taking the work-light function without entering the market. Milwaukee, DeWalt, Makita and Ryobi all sell lights designed to run on the battery the buyer already owns. Those units are counted in the tool, not here, and they win the contractor’s work-light purchase on the only argument that matters to a contractor – one charger, one battery, one platform.

Where the competition actually happens

Three contests, in descending order of value.

The agency contest is a specification contest. It is decided by whether a light survives a drop test, holds output for a rated runtime, and comes with a warranty a department can put in a procurement file. Margin here is defended by certification and by the cost of switching a fleet, and it is the only part of this market where the tariff is survivable, because the buyer is not comparing prices across a marketplace.

The lumen-per-dollar contest is the volume tier, and it is a losing contest for anyone with a US cost base. A shopper comparing two listings sees output, cell format, runtime and price. They do not see the duty paid, and the listing that paid 25 per cent on a finished import is structurally advantaged by 37.5 percentage points over the one that imported parts and paid 62.5 per cent. The contest is not between brands; it is between customs classifications.

The channel contest is the quiet one. A brand that owns a listing, a review base and a reorder habit has an asset that does not appear on any balance sheet in this sector and is not reproducible by a competitor with a cheaper bill of materials. This is the same structure visible in other US household durables – see the global household cleaning tools and supplies market, where a near-identical product and a near-identical channel produce a near-identical competitive shape – and it is why the incumbents’ real defence is distribution rather than engineering.

What the landscape will look like by 2033

Not consolidated, and slightly more bifurcated. The professional tier holds its prices because its buyers are not price-shoppers and because safety certification is a real barrier. The volume tier deflates, because the three forces in this report all push in the same direction there: a tariff that is passed through, a product that no longer generates a repeat purchase, and a channel that publishes every alternative price simultaneously.

The realistic outcome is that the seven named suppliers keep their positions, the professional tier grows its share of value without growing its share of units, and the volume tier continues to migrate to brands that exist only as a listing. The competitive risk in this market is not that a competitor wins. It is that the middle – brands large enough to carry US overheads and small enough to have no pricing power – stops being a viable place to stand.

Company references. The figures and product scope in this chapter are taken from the participants’ own published material: Energizer Holdings for the fiscal 2025 segment disclosure, Milwaukee Tool for the platform lighting range, and Princeton Tec for the tactical and industrial lines.

Chaîne de valeur

The chain, and where the value actually sits

A modern LED flashlight contains six things: an LED and its optic, a driver board, a cell or a cell compartment, a machined or moulded body with a reflector or TIR lens, a switch and sealing, and a brand. The chain that produces them runs from a handful of LED makers – Nichia, Cree, Lumileds, Samsung, Luminus – through driver and PCB suppliers in Shenzhen and Dongguan, through cell makers in China, Korea and Japan, through aluminium extrusion, machining and anodising, into assembly, and then into either a brand’s packaging or a marketplace seller’s poly bag.

What the chain is worth at each stage is the reason the tariff inversion matters so much. In a mid-market rechargeable light the LED and optic are a modest fraction of material cost, the cell is the largest single line, the body and machining are the next, and assembly labour is small enough that it is almost never the deciding factor in where a light is made. Assembly is a low-value, low-skill, easily relocated step. It is also the step the US tariff schedule charges 37.5 percentage points extra to perform in the United States.

The inversion, stated as a purchasing decision

Three ways to supply the US market with a rechargeable light, and what each pays in duty:

There is no route on that list that rewards US assembly. The only route that avoids the parts rate entirely is the one that also avoids US labour, US rent and US inventory – which is why the domestic assembly base has not grown, and why the US firms that do assemble here are explicit that their components come from everywhere.

Scale it. On an illustrative light landing at USD 12.00 with a USD 6.00 parts content and a USD 4.00 cell, the imported-finished route pays USD 3.00. The assemble-domestically route pays 62.5 per cent on the parts, USD 3.75, plus 53.4 per cent on the cells, USD 2.14 – USD 5.89, before the labour that the whole exercise was meant to create. The tariff does not close a 2.89 dollar gap; it opens one.

Upstream: the cell is the constrained node

The cell is where this chain is least substitutable and most exposed. A 21700 cylindrical cell is a globally traded commodity with a handful of credible makers, and the anode and cathode materials behind it are concentrated further still. The graphite that becomes the anode and the nickel that becomes the cathode are the two inputs whose availability and price a flashlight brand cannot influence and must simply absorb. A brand choosing a rechargeable architecture is therefore choosing to depend on that chain for the life of the product, in exchange for removing a battery from the customer’s shopping list. The trade is usually worth it on margin per unit and always worse on supply risk.

Downstream: the chain now ends at a listing

The last node has changed more than any other. A decade ago a light reached a buyer through a distributor, a rep and a shelf; today a meaningful share reaches them through a marketplace where the brand is a listing, the packaging is a photograph and the shelf position is a ranking. That removes two margins from the chain and transfers the information advantage to the platform. The global e-commerce market is the structural context; flashlights are an unusually pure case of it, because the product’s entire specification is four numbers and four numbers are exactly what a comparison table displays.

What the chain cannot do

It cannot move assembly to the United States while the parts rate exceeds the finished rate. It cannot de-China the cell quickly without paying a 53.4 per cent premium on the substitute. And it cannot pass the whole duty increase to a consumer who is comparing prices in a marketplace – so the duty is absorbed somewhere between the importer’s gross margin and the retailer’s, which is to say it comes out of the two parties in the chain with the least bargaining power.

That is the value chain’s verdict on the three forces: the tariff is not a cost added at the border, it is a cost that lands on whichever node cannot move. In this chain, that is the brand.

Activité d'investissement

Nobody is building

The defining fact about investment in US portable lighting is the absence of it. No participant in this market has announced a material expansion of US assembly capacity in this cycle, and the tariff schedule explains why: a firm that builds a US assembly line acquires a permanent 37.5 percentage point duty disadvantage on its inputs relative to the competitor importing the finished light. Capital does not need a forecast to see that. It needs only a customs classification.

Capital intensity in this category is low. A light is a design, a bill of materials, a certification file and a warranty – not a plant. The asset that matters is the brand and the listing, both of which are bought more cheaply than they are built, and neither of which appears as property, plant or equipment.

What the listed participant is actually spending on

Energizer is the only participant in this perimeter whose numbers are public, and its fiscal 2025 disclosure is instructive precisely because of what it does not contain. The company extended Project Momentum, its cost programme, into a fourth year with an explicit focus on ongoing tariff mitigation, increasing operational efficiency and integrating the APS business. Those are three ways of saying: no new capacity, no new geography, and a cost line rather than a capital line.

The year’s growth came from acquisition rather than from the category. Net sales rose 2.3 per cent to USD 2,952.7 million, of which acquisition contributed USD 63.6 million – 2.2 points of the increase – against organic growth of 0.7 per cent. In the fourth quarter, organic sales fell 2.2 per cent on a 2.9 per cent volume decline, with price contributing 0.7 points. Management described the opening quarter of fiscal 2026 as more heavily affected by temporary tariff costs. In a business where the tariff is described as temporary and the cost programme is extended to absorb it, the investment thesis is working capital and pricing, not manufacturing.

The three places money is genuinely moving

• Origin diversification. The measurable capital flow in this market is the relocation of sourcing, not of factories. Vietnam’s imports into the US grew by USD 43 million in 2025 and Thailand’s by USD 23.5 million, on a base where China still holds 80.26 per cent. That is a re-tooling of supply relationships, paid for in qualification costs and inventory, and it is the single largest investment the category is making.
• Inventory ahead of rates. When the legal basis of a duty changes three times in eighteen months, the rational response is to hold more stock than the balance sheet wants. Front-loading is a real use of cash, it flatters import volumes in the months before a rate change, and it is invisible in every published series.
• Duty as a receivable. The Supreme Court’s February 2026 ruling opened a refund channel for IEEPA duties already paid, and US Customs has stood up a processing system for claims. For an importer that brought in USD 618 million of Chinese lighting, the two IEEPA layers alone are a five-figure-million receivable, and it is worth more to a well-documented importer than to a casual one. That is a genuine asset on some balance sheets in this sector and nothing on others.

What an investor should watch

Three signals, in order.

The parts rate. If flashlight parts came off the Section 301 list, the economics of US assembly change by 37.5 percentage points in a single Federal Register notice, and this market becomes investable as a manufacturing story for the first time in a decade. Nothing else in the policy space is remotely as consequential.

Whether a marketplace-native brand is acquired. The volume tier in this market is now owned by listings rather than by factories. The first acquisition of a listing-led brand by an incumbent will mark the point at which the incumbents accept that distribution, not product, is the asset.

Whether the category’s value growth is price or volume. The forecast in this report assumes 2.7 per cent of price and mix against 1.0 per cent of volume. If the public participant’s price contribution keeps running at seven-tenths of a point while volume is negative, the price assumption is too generous and the whole forecast is high.

The honest summary for an investor is that this is a category with a real margin, a real installed base and no growth, where the return comes from holding the classification advantage and the listing rather than from building anything.

Technologie et innovation

What actually changed, and what merely got cheaper

The industry’s own framing of its history is that LEDs transformed flashlights. What LEDs transformed was the cost of light. Efficacy gains since the mid-2010s have been incremental; the dramatic shift was the replacement of the incandescent bulb, and it is over. A 2026 flashlight is not meaningfully brighter per watt than a 2020 one. It is cheaper per lumen, smaller, and more likely to be sealed and rechargeable.

That distinction matters for the forecast. A technology story based on breakthroughs implies unit growth; a technology story based on cost deflation implies value growth below unit growth, which is what the 1.0 per cent volume assumption in this report reflects.

Four architectural choices that decide the economics

• Cell format. The 18650 has been the volume standard for a decade and the 21700 is displacing it in higher-output lights because it carries more energy in a package that still fits a hand. The choice is now a supply decision as much as an engineering one: the cell that is easiest to source at a tolerable duty is not always the cell that gives the best runtime curve.
• Integrated or removable cell. An integrated cell allows a sealed, waterproof, USB-C body and a thinner product; a removable cell allows the customer to carry a spare. The first is better for the brand’s gross margin on the unit and worse for the customer’s total cost; the second preserves the battery aftermarket. Almost every premium light has moved to integrated, and that is the single most consequential decision in the category – it is what converts a razor-and-blade business into a durable good.
• USB-C and the death of the charger. USB-C PD has removed the proprietary charging cradle and the wall wart from the bill of materials and from the customer’s drawer. It has also removed the brand’s ability to lock a customer into an accessory line. The light is now charged by the same cable as the phone.
• The driver, not the emitter. In high-output lights the differentiator is thermal and driver engineering – how long the light holds its top output before stepping down, and how honestly it reports that. This is where a serious brand can still differentiate against a marketplace listing, and it is invisible in the four numbers a comparison table shows.

The standard that was supposed to settle arguments

Portable lighting has a performance standard – ANSI/NEMA FL1, maintained by the Portable Lights American Trade Organization – which defines how runtime, beam distance, peak beam intensity, impact resistance and water resistance are to be measured and reported. It is voluntary, it is self-certified, and it is the reason a product page can say 2,000 lumens without anyone having checked. A standard that is self-declared and unenforced does not discipline a marketplace; it gives a marketplace a vocabulary. The brands that benefit from FL1 are the ones whose numbers would survive a test, which is the professional tier, and they are not the ones the volume tier is buying.

The blind spot: lights that live inside tools

The most important technology shift in this market is happening outside its perimeter. Cordless-tool platforms have absorbed the work-light function: a contractor buying a light now buys a bare tool that runs on the batteries already in the van. Those lights are excellent, they are effectively subsidised by the battery ecosystem, and they do not appear anywhere in a portable-lighting trade series because they are classified and sold as tool accessories.

This is why the perimeter in this report is stated so precisely and why the technology chapter carries the caveat. The measurable market is the residue of the category after the tool platforms have taken the part of it that attaches to a battery the customer already owns. That residue is not shrinking in value – the professional and emergency segments are real and growing – but it is a residue, and a forecast that treats it as the whole category is measuring the wrong thing with great precision.

Patent activity, honestly

Patent intensity in this category is low to moderate and concentrated in mechanical and thermal details – reflector geometry, tailcap switching, mounting interfaces, driver topologies – rather than in the emitter, which is owned upstream. There is an active community of enthusiast firmware for programmable lights, which has produced genuine interface innovation that commercial brands have adopted, and there is very little that a competitor cannot design around within a product cycle. The moat in portable lighting has never been the patent portfolio. It is the certification, the listing and the warranty.

Risque de marché

The five risks that matter, ranked

1. Tariff legal churn. The rate on a Chinese flashlight has had three legal bases in eighteen months: IEEPA from February 2025, Section 122 from 24 February 2026, and Section 301 from 24 July 2026. Two of those three are gone – one struck down by the Supreme Court, one expired by statute after 150 days. The current basis, Section 301, is durable in a way the others were not, because it is a delegation of tariff authority with a statutory procedure behind it. But durable is not fixed: the rate, and the product list it applies to, are administratively revisable, and the refund channel opened on the earlier layers is still live. A buyer planning a twelve-month purchase cannot price this risk, only absorb it.

2. Channel deflation. E-commerce was 17.1 per cent of US retail sales in the second quarter of 2026 and grew 12.2 per cent year on year against 6.7 per cent for retail overall. In a category whose specification is four numbers, a channel built on comparison tables does not merely grow; it compresses the price of every listing toward the cheapest credible one. A tariff pass-through in this channel does not raise the shelf; it re-ranks it.

3. Substitution from both ends. At the low end the phone has already replaced the household torch for most of the occasions a household torch was for. At the work-light end the cordless-tool platforms have absorbed the function entirely, at a price that is subsidised by a battery the customer already owns. Both substitutions are invisible in this market’s trade data and both reduce the addressable perimeter without reducing any published number.

4. Origin concentration. China supplied 80.26 per cent of 2025 imports. Vietnam, Thailand and Cambodia together supplied 15.19 per cent, and their growth is the category’s principal mitigation. That is still a supply base where one origin is four-fifths of the product and where a port disruption, an export control or a single bilateral event moves the entire market’s landed cost. The diversification is real and it is two years old; it is not yet a hedge.

5. Demand lumpiness. A meaningful share of household demand is triggered by weather. Emergency-preparedness buying spikes before a named storm and collapses after it, which means one quarter’s import surge can be a forecast error rather than a signal. Any month-level series in this category – including the August 2026 import figure cited in this report – should be read with a seasonality caveat attached.

What is not a risk, despite appearances

Demand. The installed base is enormous, replacement is non-discretionary in the professional tier, and there is no plausible scenario in which US buyers stop needing portable light. This is not a category with a demand problem; it is a category with a margin problem.

Competition from imports. Competition from imports is not a risk to this market; it is the market. The risk is the reverse – that the tariff structure makes the imported finished product so relatively advantaged that the domestic node becomes purely a brand and a warranty, with no operational substance left.

The risk this report does not model

A change in the classification of a rechargeable light. A sealed light with an integrated lithium-ion cell is currently one article under 8513.10.20 at 25 per cent. If a future action or ruling treated the integrated cell as separately dutiable – which is a classification question, not a policy one – the effective rate on the fastest-growing part of this category would jump toward the cell’s 53.4 per cent. Nothing in the current schedule does this, and no announced action proposes it. It is listed here because it is the one plausible event that would move this forecast by more than the 3.0 per cent sensitivity band, and because classification risk in this category has historically been decided case by case rather than by rule.

Paysage réglementaire

Two regulatory regimes stacked on one product

A flashlight entering the United States is regulated twice: as an imported article under trade law, and as a consumer product under safety law. The first is where the money is and the second is where the liability is, and both have been moving.

Trade law, which now governs the business

The relevant authorities and their current status:

• Section 301 of the Trade Act of 1974. The operative authority today. A statutory delegation with an investigation and comment procedure behind it, which is why it survived the ruling that destroyed the IEEPA tariffs. The current 12.5 per cent action on Chinese goods comes from a determination covering 60 economies investigated for failing to impose and enforce a forced-labour import prohibition – 54 of them, including China, were found not to impose one at all, and 12.5 per cent is the rate applied to those. It took effect on 24 July 2026.
• Section 232 of the Trade Expansion Act of 1962. Sectoral, and the reason flashlight parts cost what they do: the metals-derivative action applies to articles of aluminium and steel, and a machined aluminium body is an article of aluminium. It is why the parts line totals 62.5 per cent rather than 37.5.
• Section 122 of the Trade Act of 1974. A 150-day bridge statute capped at 15 per cent, used from 24 February 2026 and expired by its own terms. Not currently available without a new proclamation.
• IEEPA. No longer a tariff authority. The Supreme Court held on 20 February 2026 that the statute’s power to regulate importation does not include the power to tax, and both the fentanyl and reciprocal programmes fell. The administration has continued to use IEEPA to suspend the de minimis exemption, which is not a tariff and is being litigated separately.

The practical regulatory burden of this regime is not the rate; it is the classification. The difference between 25 per cent and 62.5 per cent on the same physical product turns on which ten-digit line an entry is filed under, and the rulings history in this heading is dense with disputes about whether a given article is a flashlight, a lantern, a vehicle light, a toy, or a part. A product development decision and a customs broker’s judgement jointly determine the largest single cost line in this market.

Product safety and battery regulation

The consumer-product side is more stable and less expensive.

• Safety listing. A mains-connected charger and, in practice, a lithium-ion cell inside a consumer device are expected to carry third-party listing – UL or ETL – against cell and pack standards. Retailers enforce this more reliably than any regulator does.
• Transport. Lithium cells are dangerous goods, and UN 38.3 testing is a precondition for shipping a rechargeable light at all. It is a fixed cost per cell design rather than per product, which favours brands that reuse a small number of cell platforms.
• Chemical disclosure. California’s Proposition 65 applies to the solder, coatings and plastics in a light and is the reason US packaging carries warnings that the same product sold elsewhere does not.
• Battery stewardship. State-level battery extended-producer-responsibility regimes are spreading, and for a brand selling a sealed light with a non-replaceable cell they raise a design question rather than a labelling one: a product the customer cannot open is a product the producer owns at end of life.
• Export markets. The EU Battery Regulation applies to any US brand selling into Europe, and its requirements on carbon footprint declaration, recycled content and removability of cells are stricter than anything in US law. For a US brand with European distribution, the EU rule effectively sets the product architecture, and the US market inherits it.

The compliance asymmetry

Here is the regulatory fact that shapes the competitive landscape more than any single rule. A brand importing a finished, certified, listed light has a modest and predictable compliance cost: one file, one listing, one transport test, one label set. A brand assembling in the United States buys the same components and then owns the assembly-level certification, the process controls, the traceability and the liability for a product it built. The tariff already charges it 37.5 percentage points more for the inputs; the regulatory regime charges it a second time for the privilege of being the manufacturer.

That is not an argument that regulation is badly designed. It is an observation about where the regime’s incentives point, and they point in the same direction as the tariff: assemble offshore, brand onshore, and let the certification travel with the imported article. Any policy that intends to reverse the decline of US portable-lighting manufacturing has to address the parts rate and the assembly-level compliance burden together. Addressing either alone will not move a single production line.