Chemical manufacturing

Unlike most of heavy manufacturing in this research, basic chemicals can be a very good business for the people already in it. Chemtrade — sulphuric acid, water-treatment chemicals, chlor-alkali and sodium chlorate, made at plants across Canada and the US — reported record 2025 revenue of C$1,997.8M, up 11.8%, and record adjusted EBITDA of C$507.4M, roughly a quarter of sales [A]. Demand is dull and durable: municipalities must treat water, pulp mills and refiners must buy acid. So the pre-screen's cut is tested against a healthy incumbent, not a sick one, and it holds — for a specific reason. Everything that produces that margin is sunk and permitted long before a tonne is sold: an acid or chlor-alkali plant, its environmental approvals, rail cars and terminals for hazardous goods, and customers on multi-year supply contracts. Even maintenance is plant-scale — the biennial turnaround of one facility, North Vancouver, cost Chemtrade about C$17.9M of EBITDA in 2024 [A]. The incumbent itself does not build to grow; it spent US$180M in 2025 buying Polytec and Thatcher assets [A]. At the petrochemical end the numbers leave ordinary scale entirely: Dow expects capital spending on its Fort Saskatchewan ethylene and derivatives complex to run at about US$1.5B a year through 2030, and Dow, with every advantage, delayed the project two years in 2025 while cutting its capital budget [A]. The establishment count looks more open than this — 188 of 322 have fewer than twenty employees [A] — but those are mostly gas fill plants, blending sites and small specialty units attached to larger networks; this screen did not examine them individually and does not claim none is enterable.

Polymerisation — ethylene into polyethylene, styrene into polystyrene — is where Alberta's cheap gas liquids become a product the world buys, and Canada's resin plants sit on some of the lowest-cost feedstock anywhere. That advantage belongs to whoever owns the cracker. The telling phrase is in Dow's own results: its Packaging & Specialty Plastics segment, one of the world's largest polyethylene businesses, had 2025 net sales of US$19,970M, down 8%, and operating EBIT of US$827M, down from US$2,373M, and the company names 'lower integrated margins' first among the causes [A]. Resin is not priced as a stand-alone product with its own margin; the producer earns the spread from gas to pellet across a chain it owns end to end, and in a year of global oversupply that spread lost two-thirds of its value even for the low-cost player. An entrant who polymerises without owning the monomer buys its raw material from its competitor and keeps only the thinnest slice of the chain; an entrant who builds the whole chain is building the multi-billion-dollar complex described in the Basic Chemical Plant record. The US data shows what kind of employer this is: 1,523 establishments with 111,780 employees at an average payroll above US$90,000 a head [A] — large continuous-process sites. Canada has only 102 establishments, three of them with five hundred or more employees [A]. Just over half — 53 — have fewer than twenty employees, and those are likely compounders, specialty-resin formulators and recyclers rather than polymer plants. That end is plausibly enterable and this screen did not examine it: no anchor at that scale was found, and the record does not pretend otherwise.

Nitrogen and phosphate synthesis is world-scale plant and was never the question. The reachable end of this group is formulation: buy technical-grade active ingredient or fertilizer materials, blend, package and sell under a label. The Canadian shape says people do exactly that — of 234 establishments, 71% have fewer than twenty employees [A]. The cut is that in this industry the label is the asset, and the label is a registration. No pest control product can be sold in Canada without a Pest Management Regulatory Agency registration, product by product and use by use, and the data package behind it is sunk before the first litre ships; fertilizers and supplements answer to the CFIA under the Fertilizers Act. American Vanguard — a listed formulator and registrant of exactly this kind — shows what holding them costs: $11.3M of registration expense in 2025 on $515.1M of net sales, which fell 6%, for a net loss of $49.9M after a $126.3M loss in 2024 [A]. Its own filing names the second problem: generic competitors buy their way onto the same data through compensation offers, source from India and China, and run on thinner margins. So the entrant pays the regulatory toll in full and still does not own a moat once it is through the gate. Blending plain fertilizer avoids most of the registration burden, but that is a farm-retail business tied to the big distributors' networks and was not screened here. The software sold to this branch is screened separately.

Paint is not hard to make. A dispersion mill, tanks and a filling line are modest plant by chemical standards, resins and pigments are bought in, and Canada still has 263 establishments in the group, 52% of them with fewer than ten employees [A]. The cut is the channel. Sherwin-Williams closed 2025 at $23.57B of net sales, up 2.1%, at a 48.8% gross margin [A], and the way it earns that margin is the point: it sells through 4,853 company-operated paint stores across the United States, Canada and the Caribbean, and added 80 net new ones in the year [A]. The painting contractor — the buyer who matters in architectural paint — opens an account at a store a few minutes from the job, gets tinted product the same morning and is billed monthly. The manufacturer owns that counter. What is left to everyone else is the home-centre shelf, which a few national chains control, and the independent dealer, who already carries a national line. A new maker therefore has a product and nowhere to put it; winning a shelf means funding the tint machines, the colour system and the co-op advertising that come with it. And the category gives no help: Sherwin-Williams' store sales rose 3.2% on price, with volume down, so share has to be taken rather than found. Industrial and specialty coatings, where the sale is a specification rather than a shelf, may be a different answer and were not examined. Adhesives were not examined either. The software sold to this branch is screened separately.

Contract manufacture makes the product side of a cleaning or personal-care brand genuinely accessible — a formulator will run a batch without a factory. The cut is the category's own growth rate. Church & Dwight closed 2025 at $6,203.2M of net sales, up 1.6%, with organic growth of just 0.7% [A]. Clorox closed fiscal 2025 at about $7.10B, up 0.16% [B]. These are the scale players, and the category is flat. A flat category means share is taken, not created — and it is defended at a 45.2% adjusted gross margin [A] by companies that can meet a price cut and outspend any new brand on shelf. The entrant's real product is distribution, which is the one thing contract manufacture does not supply.

This is a residual group and it cannot be screened as one market. It holds printing ink, explosives, custom compounding of purchased resins and an 'all other' tail that runs from toner to fireworks — 440 Canadian establishments, 52% with fewer than ten employees [A], sharing nothing but a classification. Explosives manufacture is a federally licensed activity sold to mines and quarries by a few global houses; it was set aside without research. The most reachable niche is custom compounding: buy resin from the petrochemical majors, add colour, fillers and additives on a twin-screw extruder, and sell pellets to moulders and extruders who need a material the resin maker will not bother to make. It is real work, the equipment is ordinary, and customers do value a local lab that will match a colour this week. The cut is what the scale player's numbers say about the category. Avient, which describes itself as the link between large chemical producers and processors of polymers, reported 2025 sales of $3.3B from 98 manufacturing sites, up 0.6% — and down 0.3% once currency is removed [A], with declines in consumer, industrial and energy end markets. A compounder is paid a spread between a resin price it does not set and a customer who re-quotes the job every year; when the volume pool is flat, the only growth is a job taken from a neighbour at a thinner spread. Printing ink rides on print volumes and was looked at no further than that. Anything attractive inside 3259 needs its own six-digit record; this one screens compounding and says plainly that the rest was not examined. The software sold to this branch is screened separately.