Transportation equipment manufacturing

Nobody screens a new light-vehicle assembly plant: the twelve Canadian establishments with 500-plus employees are the transplants of global automakers, and US County Business Patterns puts the average establishment in this group at roughly 750 people. The enterable-looking end is the one the electric transition seemed to open — a purpose-built electric school bus and medium-duty truck maker, selling into subsidised fleet programmes. Canada ran that experiment in public. Lion Electric of Saint-Jérôme delivered 852 vehicles in fiscal 2023 for $253.5M of revenue — and a gross loss of $5.5M and a net loss of $103.8M [A]. It had put $148.0M in 2022 and $72.2M in 2023 into its Joliet, Illinois plant and its Mirabel battery campus, plus $79.1M and $67.2M into vehicle and battery development [A], against an order book it valued at about $500M — an order book that, by its own definition, included vehicles for which subsidy applications had merely been filed [A]. In December 2024 it entered creditor protection owing more than $244M; a Quebec investor group bought what remained in May 2025, and the Joliet and Mirabel sites were shut [B]. The mechanism is the cut. Plant, tooling, certification and a service network are sunk years before volume arrives, and the volume is released by government funding rounds, not by the buyer's own budget — so the entrant carries automotive fixed costs at a few hundred units a year. Truck Body & Trailer Manufacturing (3362) is cut on the same factor for a cyclical reason; here the fixed cost was never covered even at the top of the order book. Upfitting and specialty conversion on a bought chassis is the reachable adjacency and sits in 3362, not here.

A plant business on a freight cycle, and the cycle is presently against it. Wabash National — the one listed pure-play — turned $1.54B in 2025, down from $1.95B in 2024, and closed the year with a 12-month backlog of $576M against $813M a year earlier and a total backlog of $705M against $1,169M [B]. That is a roughly 40% collapse in visible forward work while fleets defer capital spending. Wabash is put at about 7.0% of US truck-trailer manufacturing revenue [C], which implies an industry near $22B [derived] shared with Great Dane, Utility Trailer and Trailmobile. Entry means a welding and assembly plant, a dealer network and working capital through exactly this kind of trough. The reachable adjacency is upfitting and service bodies at small scale, which needs a shop rather than a plant, and which this record does not screen.

This is the one part of the vehicle chain with real Canadian depth: 731 establishments, 458 of them in Ontario, and 198 employing a hundred people or more — a size profile unlike almost any other manufacturing group, and the first sign that small does not survive here. Martinrea International, the Vaughan-based metal-forming and fluid-systems supplier, shows what the economics are for a well-run incumbent. In 2025 it made C$4,821.9M of sales and C$268.1M of adjusted operating income — a 5.6% margin, which was above the midpoint of its own outlook [A]. To earn that it spent C$237.7M on plant and equipment, about 4.9% of sales, across 57 locations in ten countries [A]. The mechanism behind those two numbers is the cut. The customer is one of a handful of automakers, who award a part for the life of a vehicle programme, require the supplier to build the tooling and capacity before the programme launches, and then negotiate the price down each year. Volume is whatever the automaker's assembly line runs — Martinrea's release notes it had to negotiate recovery of its tariff costs from customers rather than price them in [A]. An entrant therefore needs quality-system certification, a launch record an automaker's purchasing group will accept, and a press line or moulding cell financed ahead of revenue, in exchange for a mid-single-digit margin it does not control. Capital is the binding factor because it is committed before the award is secure. The 191 establishments under five people are largely rebuilders and aftermarket or performance-part makers, who sell to distributors and enthusiasts rather than to automakers. That is a different proposition, arguably the reachable one, and it was not examined here.

This is the one corner of aerospace where demand really has moved. AeroVironment's revenue went from $820.6M to $1,976.8M in its year to April 2026 [A], with $1.46B of funded backlog by August [A]. Ukraine's procurement agency signed UAH 333.6 billion of drone contracts in the first half of 2026, twice the year before [A], and the EU set up a €150B loan instrument for common defence buying [A]. Capital is not short either: Anduril raised $5B at $61B, Shield AI $2B, Quantum Systems $1.2B and Helsing $1.8B, all in 2026 [B]. The cut is what has to be in place before the first government sale, and none of it is money an ordinary entrant can raise. In the US a buyer is choosing from the Defense Contract Management Agency's Blue UAS list, and since December 2025 the FCC has closed the normal authorisation route to new foreign-produced drones and critical components [A, as described in Red Cat's 10-K]. In the EU, SAFE money needs at least 65% of component costs from the EU, Ukraine or the EEA [A]. In Ukraine, 80% of drone money goes to products rated on combat data, and only 20% is kept for testing new ones [A]. Add facility clearance, ITAR and CMMC, and the entrant is funding years of qualification against a budget that is mostly still a request: the Pentagon asked for $54.6B for its Defense Autonomous Warfare Group in FY27 after giving it $226M in FY26, and $53.6B of that rides on a reconciliation bill that may not pass [B]. A programme win does not end the cost. Red Cat, an Army Short Range Reconnaissance winner, lost $61.8M on $35.7M of revenue in the first half of 2026 [A]. Incumbents widen the gap by buying in: AeroVironment paid $3.48B for BlueHalo [A]. The sibling record 336410 reads the Canadian side the same way. What can be entered is the supply chain under the primes: compliant components, test services and compliance tooling.

The largest and fastest-growing market in the portfolio, and the one where enthusiasm and executable opportunity are most badly mismatched. Budget headlines are not revenue.

Thirty-one Canadian establishments, of which a handful are the whole industry — passenger rolling stock built under public transit contracts in Quebec and Ontario, and freight cars built in Hamilton. The passenger side is a procurement contest among global groups and is not an entry proposition at any ordinary scale. The freight side looks more approachable: a railcar is welded steel on bought trucks and couplers, and a lessor or railroad orders hundreds at a time. The cut is what those orders look like over a cycle. Greenbrier, one of the two large North American builders, called fiscal 2025 a record — $3,240.2M of revenue, 22,000 units delivered, an 18.7% aggregate gross margin — yet took orders for only 13,200 units, ended the year with a backlog of 16,600 units worth $2.2B, guided fiscal 2026 down to 17,500–20,500 deliveries and $2.7–3.2B of revenue, and announced two further plant closures in the fourth quarter [A]. That is a record year in which the order book ran at 60% of output. Demand is replacement of a long-lived fleet, released in lumps by a small number of lessors and Class I railroads when freight volumes and interest rates allow, and the incumbents respond by flexing plants in Mexico and owning lease fleets (Greenbrier's is 17,000 cars at 98% utilisation [A]) that carry them through the trough. An entrant has the plant but not the lease book, and meets its first downturn with one product and no annuity. The reachable work around this industry — car repair, wheel and component reconditioning, track-maintenance equipment — is mentioned in the definition and was not examined.

Shipyards are not an entry proposition — the seven Canadian establishments with 500-plus employees live on federal shipbuilding programmes. The enterable half is boat building, and the size bands show it is genuinely a small-shop trade: 186 of 304 establishments have fewer than ten people, concentrated in British Columbia and Nova Scotia. A mould, a shed and a skilled laminating or welding crew will produce a boat. The cut is what happens between the shed and the buyer. Recreational boats reach customers through independent dealers who finance their stock, so a builder's shipments are set by dealer inventory, not by retail demand — and when dealers destock, wholesale volume falls faster than the market. Malibu Boats, a well-run listed builder, shipped 4,898 units in fiscal 2025, down 9.0%, for $807.6M of net sales and a 1.9% net margin, and described the year as spent 'supporting our dealers' efforts to reduce their inventory' [A]. The more telling anchor is Canadian. BRP — with a global dealer network, its own engines and C$8,442.7M of revenue in the year to 31 January 2026 — bought its way into boats some years ago and has now left: it booked a C$277.6M net loss from discontinued marine operations in fiscal 2025, a further C$51.1M in fiscal 2026, and closed the sales of Alumacraft's and Manitou's assets during 2025 [A]. If the strongest powersports distributor in the country could not make aluminium fishing boats and pontoons pay through one downturn, a new builder selling a discretionary, financed, seasonal product through the same dealers will not. Not examined: the commercial end — fibreglass lobster boats in Nova Scotia, aluminium workboats in BC — where the buyer is a licence-holder with an income from the vessel. That is a different demand curve and deserves its own look; the marine dealer software on this branch is screened separately.

A residual code — motorcycles, snowmobiles, all-terrain and side-by-side vehicles, golf carts, bicycles, military vehicles and more. The screen takes powersports, because that is where a Canadian entrant would be looking and where the public record is, and says plainly that the rest was not examined. The attraction is that enthusiasts pay real money for a better machine and that Quebec has the supplier base and the engineers; 71 of the 191 Canadian establishments are there. The reason they are there is also the cut. BRP of Valcourt — Ski-Doo, Sea-Doo, Can-Am — reported C$8,442.7M of revenue for the year to 31 January 2026, up 6.8%, and C$1,103.4M of normalised EBITDA [A]. In its fourth quarter its North American retail sales rose 12% with market-share gains in off-road vehicles and snowmobiles, in a year-round-product industry it described as flat [A]. This is an incumbent taking share in a flat market, with its own engines, a new-model cadence it funds from a billion dollars of EBITDA, and the dealer network every entrant would need. The places an entrant might hope to outflank it are electric and light-mobility products — and BRP's own results close that door too: it recorded a C$229.8M impairment on its EV assets and light-mobility unit and C$28.5M of EV-related provisions, citing 'the challenges in the EV industry' [A]. The best-resourced player tried the new category and wrote it down. There is no slow or distracted incumbent here to be vulnerable. The small establishments in this code are makers of things like custom motorcycles, bicycle frames and specialty carts; each is its own niche and none was screened.