Pipeline transportation

A crude line is the best business in the oil patch once it exists: shippers sign long contracts, the toll is approved rather than negotiated, and the barrels have nowhere else to go. Enbridge's Liquids Pipelines segment earned $9,710M of adjusted EBITDA in 2025 on a Mainline that averaged 3.1 million barrels a day and was apportioned for nine months of the year — demand exceeded the pipe [A]. That is the attraction, and it is also the trap. The only new Canadian export line built in a generation, the Trans Mountain expansion, saw its cost estimate rise from $21.4B to $34.2B, and the Parliamentary Budget Officer values the finished asset at $29.6B to $33.4B depending on whether contracts renew [A]. A completed, full, tolled pipeline is worth less than it cost to build, and the owner that absorbed the difference was the federal treasury. Meanwhile the incumbent adds capacity by optimising steel already in the ground: Enbridge sanctioned 150,000 barrels a day of new Mainline capacity for US$1.4B [A], a cost per barrel no greenfield route can approach. The 38 Canadian establishments are not a fragmented field — six employ more than 500 people and the small ones are field offices and feeder systems tied to the same few owners. Regulation is a second wall, but it is not the binding one: a fully permitted line still has to be financed against an incumbent whose next barrel of capacity is nearly free. The measurement and accounting software sold to these operators is screened separately at 486.

Gas transmission looks like crude transmission and is cut for a different reason. The crude record at 4861 turns on what a new line costs; this one turns on who is allowed to spread that cost. TC Energy's Canadian Natural Gas Pipelines segment earned $3,687M of comparable EBITDA in 2025, up from $3,388M, on an NGTL System whose receipts averaged 15.5 Bcf a day [A]. NGTL is a cost-of-service franchise: when it adds a lateral to a new producing area, the capital goes into a rate base that every existing shipper pays for. An entrant proposing the same lateral must recover the whole cost from the new shippers alone, so the incumbent's toll for identical steel is lower by construction, and the regulator that approves both has no reason to prefer the dearer one. The company spent $6,337M of capital in 2025 extending systems it already owns [A]. The one large greenfield line of the decade shows what is left over: Coastal GasLink, about 670 km built to feed a single LNG terminal, reached an estimated cost of approximately $14.5B, with the sponsor citing labour shortages and contractor underperformance [A] — and it was the incumbent that built it. Of 46 Canadian establishments, 25 have fewer than five employees; those are not challengers but single-asset entities and producer-owned connections. There is no weak incumbent here to take share from. The measurement and accounting software on this branch is screened separately at 486.

What is left in the pipeline subsector after crude and gas: lines carrying gasoline, diesel and jet fuel from refineries to terminals, and an 'all other' tail of carbon dioxide, slurry and similar single-purpose lines. The asset is attractive for the same reason the trunk lines are — once built it is the cheapest way to move the product and nobody builds a second one beside it. The cut is that the number of corridors is fixed by the number of refineries, and nobody is building refineries. The clearest price signal is American: ONEOK bought Magellan Midstream, owner of the longest refined products pipeline system in the United States with access to nearly half the country's refining capacity, in a transaction valued at approximately $18.8B including assumed debt, paying a 22% premium [A]. That is how capacity changes hands in this industry — a gas company buying an existing network whole. Two years on, ONEOK's FY2025 10-K shows what it bought as a working segment: Refined Products and Crude revenue of US$13,039M and segment adjusted EBITDA of US$2,177M, on 9,800 miles of refined products pipelines, 53 products terminals and 100 MMBbl of storage [A]. The one place new pipe is being laid is inside that system — ONEOK reports it is building a 230-mile line from Scott City, Kansas to the Denver airport to add 35 MBbl/d of capacity to its own network [A] — an extension by the corridor's owner, not a second corridor beside it. In Canada the group is 44 establishments, 29 of them with fewer than five employees and one above 500. That shape is a handful of real systems, generally owned by or contracted to the refiners that ship on them, plus small legal entities holding a single line. An entrant has no route to a first customer: the shipper already owns the pipe or is contracted to it, and product demand gives no reason to add a corridor. The specialty tail — carbon dioxide lines for sequestration are the live example — is policy-created, project-financed and was not examined here. The software sold into this branch is screened separately at 486.