Bitumen does not flow. To move it by pipeline, producers blend it with a light diluent, mostly condensate, until about 30% of the barrel is diluent. Alberta used 781,000 b/d of pentanes plus in 2025 and produced 425,000 b/d. The other 357,000 b/d came from BC and the US. Verified

Canada is now planning the pipes to grow the oil sands: Trans Mountain and Enbridge expansions, South Bow’s revival of Keystone XL, and Pacific Link, which Ottawa named its first national interest project on 1 October. Verified None of them brings diluent back. We built the full balance and our own year-by-year forecast, including every sanctioned LNG plant and Ksi Lisims, and set out what it would take to bring the diluent in, who pays, and who can act.

In 60 seconds

  • The gap is growing on its own: 357 kb/d today, 430 kb/d by 2031 on the AER’s forecast, which assumes less than half the past decade’s oil sands growth.
  • Every new oil export pipe is a diluent customer. About 1.1 mb/d could open by 2030, and the AER fills only about half of it; Pacific Link alone would need 300 kb/d.
  • With every oil export pipe full and all LNG projects running, Ksi Lisims included, Alberta would need about 555 kb/d of diluent from outside, 55% more than today: about C$18 billion a year at 2025 prices, against C$11.5 billion now.
  • Two waves. At the historical growth pace, Alberta needs about 70 kb/d of diluent beyond today’s import capacity in 2029, before the LNG gas flows, and about 150 kb/d by 2037 as Pacific Link fills.
  • Gas helps only if it is rich. Even on the AER’s generous yield, closing the gap with gas takes about two more LNG Canada-sized plants beyond Ksi Lisims, C$65 billion or more of investment; at Alberta’s current average ratio, almost five.
  • No single fix, and none is easy. Rail, diluent recovery and splitters can bridge to 2030; a lasting fix needs pipe that brings diluent north, and producers must sign for it. If nothing is built, price rations diluent, growth slows, and Pacific Link, the last to open, runs part-full.
  • Who moves first. Enbridge holds the fastest pipe fix, Shell the biggest supply lever, and SAGD producers the contracts that make either bankable. Our six calls are in Section IX.

Alberta Already Brings In 357,000 Barrels a Day

Alberta’s condensate output grew 7.2% in 2025 on liquids-rich drilling, but demand grew from a much larger base. The AER sees the shortfall reaching 430 kb/d in 2031 and staying above 410 kb/d to 2035. Verified Most of the gap comes from the US through two pipelines from the Chicago area that run at or near capacity; the rest comes from BC’s Montney and by rail. Reported At 2025 prices, that is about C$11.5 billion a year of diluent bought from outside Alberta. Estimate

Fig. 1The gap was 370 kb/d in 2015; the AER sees 430 kb/d in 2031
Alberta pentanes plus supply and demand, kb/d, 2015–2035
What it shows → Alberta made 425 kb/d of pentanes plus in 2025 and used 781 kb/d, mostly to dilute bitumen. The AER sees the shortfall rising to 430 kb/d by 2031.
How to read it → Red line: demand. Green line: Alberta’s own supply. The shaded area is the gap that must be brought in; the grey panel is the AER’s forecast.
Why it matters → Demand follows bitumen output; supply follows liquids-rich drilling, which grows more slowly. The gap widens even before the new pipes.
AER forecast0200400600800100020152020202520302035gap 370gap 357gap 430Demand for diluent: 971Alberta supply: 555Thousand barrels a day (kb/d)The gap is brought in from BC and the US by pipeline and rail.
Source: Alberta Energy Regulator, ST98-2026, Figure S6.4 (2025 estimated; 2026 onward forecast).
Table 1Where Alberta’s diluent comes from (2025, kb/d)
Sourcekb/dHow it gets there
Alberta’s own pentanes plus425Montney and Duvernay gas plants, by pipeline to Edmonton and Fort Saskatchewan
United States, by pipeline260–290Two lines from the Chicago area, supplied from the US Midwest and Gulf Coast; running at or near capacity
BC and rail~65–95The rest of the gap: BC Montney condensate by pipeline, and US barrels by rail
Total pentanes plus used781Mostly blended into non-upgraded bitumen
Butane blended~124Alberta’s own LPG, blended up to pipeline vapour-pressure limits, mostly because it is cheaper than crude
All diluents, Western Canada~900Condensate and other light hydrocarbons
Sources: AER ST98-2026 (Figures S6.3, S6.4; Table S8.6); CER, Line 13 expansion; Rystad Energy; S&P Global via The Hub (14 Oct 2026). US range: Rystad’s estimate of imports (about 260) to the two lines’ capacity (about 290). BC and rail is an Insight Stream estimate: the AER gap less US pipeline imports. Why butane is blended is our interpretation.

Every New Oil Export Pipe Is Also a Diluent Customer

About 1.1 mb/d of new oil export pipe could open by 2030: Trans Mountain’s optimization (420 kb/d), Enbridge’s first Mainline phase (150 kb/d), Express-Platte (30 kb/d) and, if it proceeds, South Bow’s Prairie Connector, 465 kb/d at the end of 2028 and expandable beyond 800 kb/d with more pumping. Reported Enbridge postponed its 250 kb/d second phase in August for lack of producer commitments. Reported Pacific Link, 1 mb/d to Roberts Bank near Vancouver, would follow in the early 2030s. Filled with diluted bitumen, every 1 mb/d of pipe needs about 300 kb/d of diluent, about C$10 billion a year at 2025 prices. Estimate

Table 2The new oil export pipes, and the diluent each would need if full
Pipelinekb/dStartStatusIn AER forecast?Diluent if full, kb/d
Express-Platte expansion302026Under wayYes9
Enbridge Mainline, phase 11502027PlannedYes45
Enbridge Mainline, phase 2250PostponedDeferred Aug 2026 for lack of producer commitmentsYes75
Trans Mountain optimization, incl. Puget Sound4202030PlannedYes126
Prairie Connector (Keystone XL revival)465End-2028 (company target)Expandable to 800+; 465 kb/d of 20-year commitments; FID mid-2027; needs a durable permitNo140–240
Pacific Link1,000Early 2030sNational interest project; construction targeted from Sep 2027; not yet approvedNo300
Sources: AER ST98-2026, Table S8.7; Pipeline & Gas Journal (29 May 2026); S&P Global (1 Oct 2026); The Hub (1 Oct 2026). Diluent at a 30% blend, if the pipe fills with new diluted bitumen; less if part carries synthetic or conventional crude. Pacific Link’s date is an Insight Stream estimate. South Bow Q2 2026 call (Prairie Connector capacity and timing); Reuters via Pipeline & Gas Journal (3 Aug 2026; Enbridge phase 2).

How fast they fill depends on growth. The AER’s forecast adds about 46 kb/d of bitumen a year to 2035 and fills about 0.55 mb/d of new pipe by 2031. Over the past decade the oil sands grew about 120 kb/d a year; at that pace the pipe ahead of Pacific Link fills by about 2031, close to Enverus’s view, and Pacific Link by about 2037. Estimate And all of the planned growth is bitumen that needs diluent, as no material upgrader expansions are planned. Interpretation

Fig. 2At the past decade’s pace, every pipe fills by the late 2030s; on the AER’s path, they stay part-empty
Alberta bitumen production, mb/d, 2010–2038
What it shows → The oil sands grew about 120 kb/d a year over the past decade, through two price crashes. The AER’s forecast adds 46 kb/d a year.
How to read it → Navy: actual output. Grey dashes: the AER’s forecast. Gold dashes: the past decade’s pace continued. Red dotted lines: the output needed to fill the pipes; the dots mark when the gold line reaches them.
Why it matters → At the historical pace the pipes before Pacific Link fill by about 2031 and Pacific Link by about 2037. On the AER’s path, neither fills by 2035.
12345201020152020202520302035Pipes before Pacific Link full~2031Pacific Link full too~20373.5AER: +46 kb/d a yearPast decade: +120 kb/d a yearMillion barrels a day of bitumen (upgraded and non-upgraded)Fill levels assume 0.7 bbl of bitumen per barrel of new pipe. 2026 onward is the AER forecast; the dashed gold line is an Insight Stream projection.
Source: AER ST98-2026, Figure 9 (2026 onward forecast); Enverus via BOE Report (21 Apr 2026); Insight Stream projection at 120 kb/d a year from 2025.

With Every Export Pipe Full: About 555,000 Barrels a Day

We start from the AER’s 430 kb/d in 2031, add about 150 kb/d for the pre-Pacific Link pipe the AER does not count and 300 kb/d for Pacific Link, then subtract the condensate from the incremental gas production needed to meet new demand: LNG Canada’s second phase, Ksi Lisims (assumed sanctioned), Woodfibre, Cedar, Meta’s data-centre campus and fuel for the new SAGD barrels, about 5.6 Bcf/d in all. Estimate

Fig. 3Even with Ksi Lisims, full oil export pipes leave Alberta needing to bring in about 555 kb/d
Alberta diluent balance with every new oil export pipe full, kb/d, 2030s: central case and range
What it shows → With every oil export pipe full, Alberta would still need to bring in about 555 kb/d of diluent on the AER’s yield, even with Ksi Lisims: a 55% increase from today’s import volume.
How to read it → Read from top to bottom: start from today’s shortfall, add what each new oil export pipe needs, subtract the condensate from gas produced for new demand. Solid bars are the central case; whiskers show the range from the richest gas to dry gas.
Why it matters → New gas offsets only part of the new pipes, even in the central case. The rest has to be brought in.
Thousand barrels a day (kb/d) of pentanes plus. Solid: central case (AER yield). Whiskers: range.02004006008001000Today (2025)AER shortfall3572031, AER forecastalready fills ~0.55 mb/d of new pipe430+ Oil export pipes before Pacific Linkthe ~0.5 mb/d the AER does not fill+150 → 580+ Pacific Link1 mb/d of diluted bitumen at ~30%+300 → 880− Condensate from gas for new demandLNG Canada 2, Woodfibre, Cedar, Meta, SAGD fuel−230 central; −305 if rich− Ksi Lisims, if sanctioned (assumed)12 Mtpa, ~1.6 Bcf/d of feed gas−90 central; −120 if rich= Still to bring inwith every oil export pipe full~555 central455–880Central: AER yield of 58 bbl/MMcf. Range: all new gas as rich as the best Montney wells (77) to dry gas (0). Ksi Lisims assumed sanctioned.
Sources: AER ST98-2026; Insight Stream estimates. LNG feed gas at 0.132 Bcf/d per Mtpa; Meta at 0.16 Bcf/d per GW; SAGD fuel at 1.2 Mcf per barrel, the midpoint of AER Table S5.2. Ksi Lisims assumed sanctioned.

On the AER’s own yield, Alberta would still need to bring in about 555 kb/d of diluent, a 55% increase from today’s import volume and about C$6.4 billion a year more at 2025 prices. If all the new gas were as rich as the best Montney wells it would be about 455; if dry, 880. Estimate Without Pacific Link the central figure is about 305 kb/d, within today’s diluent import capacity: Pacific Link turns a manageable gap into a structural one. Interpretation

Rystad Energy sees the same shape: its high case, with an Alberta–BC pipe and further Mainline phases, implies a shortfall of about 383 kb/d on top of today’s imports; its base case is tightly balanced. Reported Our extra shortfall, about 200 kb/d, sits between the two because we count the condensate from LNG demand. Interpretation


The Squeeze Comes in Two Waves

We modelled the balance year by year to 2038. Demand follows bitumen output, on the AER’s path or at the past decade’s 120 kb/d a year, capped by the pipe actually open. Supply is the AER’s forecast plus condensate from each new gas customer at 58 barrels per MMcf (37–77 as the range), with gas for each LNG plant flowing a year early. The zero line is today’s diluent import capacity, about 357 kb/d. Estimate

Fig. 4The squeeze peaks in 2029, eases with LNG, then returns once Pacific Link fills
Insight Stream forecast: diluent Alberta must bring in beyond today’s import capacity, kb/d, 2025–2038
What it shows → On our central case, Alberta needs about 70 kb/d more diluent than today’s import capacity in 2029, falls back below that capacity once gas for LNG Canada’s second phase and Ksi Lisims flows, then needs about 150 kb/d more by 2037 if Pacific Link fills.
How to read it → Each line is the diluent Alberta must bring in above today’s ~357 kb/d. Navy: growth at the past decade’s pace with the pipes before Pacific Link. Red: the same plus Pacific Link from 2032, with the shaded band from Alberta’s average yield (top) to the richest wells (bottom). Grey dashes: the AER’s growth path. Right-hand labels convert kb/d into daily 120-car trains from Texas.
Why it matters → Timing matters as much as size: the hardest years come before the LNG gas flows, when only rail and diluent recovery can respond.
Pipes fillbefore LNGgas flows-200-1000+100+200Today’s routes full1 train/day2342025202720292031203320352037Gas for LNG Canada 2 and Ksi Lisims starts flowingWith Pacific LinkPipes before Pacific LinkAER growth pathkb/d Alberta must bring in beyond today’s ~357 kb/d of diluent import capacity (rail, recovery or new pipe)Below the line: less than Albertabrings in todayCentral case: AER yield of 58 bbl/MMcf. Shaded: Pacific Link case at 37 (top) to 77 (bottom) bbl/MMcf. Insight Stream model.
Insight Stream model built on AER ST98-2026 (Figures 9, S6.4), the pipeline timings in Table 2 and the gas demand in Section III. BC inflows are held at today’s level; BC condensate growth not tied to LNG would lower every line.

Three findings. The squeeze starts now: even on the AER’s path, Alberta needs about 20 kb/d beyond today’s capacity in 2026–28. It peaks first in 2029, when the oil export pipes fill before gas for LNG Canada’s second phase and Ksi Lisims flows: about 70 kb/d, or 90 at Alberta’s average yield, one to one and a half unit trains a day from Texas. LNG then brings relief until Pacific Link fills, and the requirement climbs back to about 75 kb/d by 2035 and 150 kb/d by 2037, or 260 at the average yield. Estimate These figures sit below the static balance because the AER expects Alberta’s own condensate output to keep rising after 2031.

Could producers outrun the squeeze by drilling for condensate? Alberta’s output rose about 24 kb/d a year over the past decade, and our central case already has it growing about 40 kb/d a year to 2035, because it adds the condensate from LNG demand. Estimate The limit is the gas: every liquids-rich well brings gas that needs a buyer, and until LNG demand arrives it has nowhere to go but an oversupplied Alberta market (Article 50). Interpretation On the AER’s slower path, LNG closes the gap after 2030; Enbridge says producers will not commit to new pipe until policy is settled. Reported That may be close: the five largest oil sands producers are negotiating a binding growth agreement with Ottawa and Alberta, and Enbridge said on 6 November that it sees “the cusp of a huge cyclical” growth wave. Reported If the agreement is signed, we expect growth nearer the historical pace. Interpretation


At Least Two More LNG Canadas, After the Pipes

Whether gas can close the gap depends on how rich the new gas is, and the forecasts disagree. Those that model LNG feed gas directly see it lean: an S&P Global illustration implies about 8 barrels of condensate per MMcf, and a 2018 CERI study of LNG Canada and Kitimat LNG about 21. Reported Alberta’s average was 37 in 2025, double 2015’s; the AER’s forecast implies 58 for each extra MMcf; the best Montney wells make about 77. Estimate Some LNG gas will be rich: Shell’s purchase of ARC Resources was reported as securing Montney feed gas for LNG Canada. Reported We use the AER’s 58 as our central case. It is the richest published forecast we found, so it flatters what gas can do. Interpretation

Fig. 5How rich will the new gas be? Forecasts range from 8 to 58 barrels per MMcf
Condensate per MMcf of new gas implied by each forecast; gas needed for 100 kb/d; condensate from one LNG Canada phase
What it shows → Forecasts of how much condensate new gas brings range from about 8 to 58 barrels per MMcf; only the best wells reach 77.
How to read it → Bars: barrels of condensate per MMcf under each forecast; the gold box is our central case. Right-hand columns: the gas needed for 100 kb/d, and what one LNG Canada phase would bring at that yield.
Why it matters → Forecasts that model LNG feed gas directly are the leanest. Our central case uses the richest published forecast, so it flatters what gas can do.
020406080Barrels of condensate per MMcf of new gasGas per 100 kb/dPer LNG CanadaS&P Global, LNG illustration (2026)implied from ~110 kb/d per 100 Mtpa812.0 Bcf/d15 kb/dCERI, LNG Canada + Kitimat LNG (2018)+64 kb/d of condensate for 3 Bcf/d214.7 Bcf/d39 kb/dAlberta average, 2025 (AER)all marketable gas372.7 Bcf/d68 kb/dAER forecast, each extra MMcf, 2025–30our central case581.7 Bcf/d107 kb/dBest liquids-rich Montney wellsa producer we analysed in Article 47771.3 Bcf/d141 kb/dForecasts that model LNG feed gas directly see it lean; the AER’s figure reflects today’s liquids-driven drilling. One LNG Canada phase = 1.85 Bcf/d.
Sources: S&P Global via The Hub (14 Oct 2026); CERI via Bloomberg and OilPrice.com (18 Jul 2018); AER ST98-2026 (Figures S5.2, S6.4); Article 47. Insight Stream calculations.

Table 3 asks how much more gas demand Alberta would need once everything already planned is running, at the AER’s yield, Alberta’s average and the LNG-specific forecast. Estimate

Table 3After every planned plant is running: the shortfall still left, and the gas to cover it
kb/d unless statedCentral: AER yield (58 bbl/MMcf)Alberta average (37)LNG-specific forecast (21)
Diluent to bring in with every oil export pipe full880880880
− Condensate from the 5.6 Bcf/d of gas produced for new demand already counted−323−207−118
= Still to bring in557673762
− What Alberta already brings in today−357−357−357
= New shortfall after those plants are running200316405
More gas demand needed to cover it, Bcf/d3.58.619.1
= LNG Canada-sized plants (1.85 Bcf/d each)1.94.610.3
Gas already counted: LNG Canada phase 2 (1.85 Bcf/d), Ksi Lisims (1.58, assumed sanctioned), Cedar (0.44), Woodfibre (0.28), Meta (0.15) and SAGD fuel for the new barrels (1.27). If all new gas matched the best Montney wells (77), less than one plant would do. Assumes today’s 357 kb/d of imports continues. Insight Stream estimates.

On the AER’s yield, closing the gap with gas takes about two more LNG Canada-sized plants beyond Ksi Lisims, to borrow a framing from Richard Masson; at Alberta’s average, almost five; on the LNG-specific forecast, more than ten. Estimate LNG Canada’s second phase, an expansion of an existing site, was announced as a C$33 billion investment. Reported Two more of that size would mean C$65 billion or more. Estimate Gas cannot close this gap on its own: producers drill liquids-rich gas only when the gas itself has a buyer, so it takes new gas demand, such as LNG, to bring the condensate, not demand for diluent alone. Interpretation

Timing is the other catch. LNG Canada’s second phase, sanctioned in September, is due around 2031, with gas flowing a year earlier; a plant sanctioned in 2027 would start around 2032. Reported Most of the new pipe could run by 2030, so the bridge years fall on imports. Interpretation

The LNG market is the last test, and it cuts both ways. Today it is tight, not glutted: the IEA said on 3 November that the Gulf disruption could keep it tight into 2028, and four North American projects, LNG Canada’s second phase among them, were sanctioned after March. Reported But the supply wave due this decade has been delayed, not cancelled, and a plant sanctioned now would start in the early 2030s, as it lands. Buyers asked to sign 20-year contracts will price that in; Canada’s answer is a short route to Asia that passes no chokepoint (Article 51). Interpretation

Fig. 6The pipes come first; most new gas demand arrives in the early 2030s
New oil pipe out of Alberta and new gas demand that brings condensate, 2026–2035
What it shows → About 1.1 mb/d of new oil export pipe could be in service by 2030. Of the new gas demand that brings condensate, only Woodfibre and Cedar start before 2030.
How to read it → Each row is one project. The dot marks the start of service; the pale bar runs from investment decision or construction to service. Solid rows are approved or sanctioned; light rows are not.
Why it matters → Pipes fill with diluted bitumen as soon as they open. The largest gas demand arrives in the early 2030s, so the next five years fall on imports.
2026202720282029203020312032203320342035Oil pipelines out of Alberta (capacity added, kb/d)Express-Platte expansion (+30)Enbridge Mainline, phase 1 (+150)Enbridge phase 2 (+250), postponedTrans Mountain optimization (+420)Prairie Connector (+465; 800+ possible)FID mid-2027Pacific Link (+1,000)construction from Sep 2027New gas demand that brings condensate (feed gas, Bcf/d)Woodfibre LNG (0.28 Bcf/d)Cedar LNG (0.44 Bcf/d)LNG Canada phase 2 (1.85 Bcf/d)FID Sep 2026Ksi Lisims LNG (1.58 Bcf/d), assumedFID targeted end-2026Solid: under way or sanctioned. Light: not yet approved, or postponed. Pacific Link and Ksi Lisims start dates are our estimates. Bars: decision or construction to start.
Sources: AER ST98-2026, Table S8.7; Pipeline & Gas Journal; S&P Global; World Oil (29 Sep 2026); Article 51. Prairie Connector’s date is South Bow’s target; Pacific Link and Ksi Lisims dates are Insight Stream estimates.

North America Has the Condensate. Alberta Needs the Route.

The US Gulf Coast has condensate to spare, and products pipelines already move it to the Chicago hubs. The bottleneck is the last 2,500 km into Alberta, where both import pipelines are full. Interpretation Seven options exist; none closes the gap on its own.

Fig. 7Seven routes, three timelines: what could bring diluent in, and when
Schematic map of the options to bring diluent into Alberta, with scale, timing and cost
What it shows → Seven ways to bring diluent in, where each comes from, how much it could carry and when.
How to read it → Numbered markers match the key. Solid line: today’s import pipes, already full. Dashed red: rail. Dotted: pipe that does not exist yet. Colour shows cost relative to a pipeline toll.
Why it matters → Rail and diluent recovery are the only bridges before 2030; the cheap, large options are pipes that need contracts now.
PacificOceanGulf of MexicoCANADAUNITED STATESEdmontonHardistyRoberts BankKitimatChicago hubsUS Gulf CoastWyoming12345671Expand the two US import lines10–20 kb/d now; ~75 with pumping · toll2Rail from the US Gulf Coast~60 kb/d per daily train · months · ~2× toll3Diluent recovery at Hardisty15–30 kb/d · 20284Rail from a West Coast terminal~25 kb/d · needs a new terminal5Splitting light crude at Edmonton~20 kb/d per 60 kb/d unit · ~20296New or reversed line north100–200 kb/d · 2030s · long contracts7Return line beside Pacific Link~190 kb/d · 2030s · if designed in nowSchematic, not to scale. Solid: existing import lines, full. Dashed red: rail. Dotted: pipe that does not exist yet. Grey dotted: Gulf Coast supply to Chicago.Colour = cost: teal, pipeline toll or lower; gold, moderate; red, about twice a pipeline toll or more.
Sources: CER (Line 13 expansion); Enbridge (2013 Southern Lights open season; Q2 2026 results); RBN Energy (Line 13 reversal, 2010); The Hub (Northern Gateway; rail); EnCana Form 6-K (2005), Prince George Citizen (2011) on Kitimat; Rystad Energy and Gibson Energy (diluent recovery). Cost: CAPP estimated rail from Alberta to the US Gulf at $15–22 a barrel against a pipeline toll of about $9 (crude, via the NEB, 2015). Scales and dates are Insight Stream estimates.

Until new pipe opens, only rail can grow large. A 120-car unit train carries about 60,000 barrels, so every 100 kb/d means almost two trains a day, at roughly twice the cost of pipeline. Reported At the 2015 cost gap, that is about $0.2–0.5 billion a year more per 100 kb/d. Estimate Rail from the West Coast has been done, through a Kitimat terminal from 2006 that is now part of the LNG Canada site, and light condensate likely falls outside the north-coast tanker moratorium; a new terminal would be needed either way. Interpretation

Two smaller bridges keep diluent in Alberta. Diluent recovery at Hardisty strips it from blend and ships the heavier crude by rail. A splitter turns light crude into roughly a third diluent, a third jet-range fuel and a third heavier oil, so its economics rest on the jet fuel; a decade ago, the Burnaby refinery met only about 40% of Vancouver airport’s needs. Reported Butane offers little extra room: it is already blended up to pipeline vapour-pressure limits. Interpretation

The lasting fix is pipe, and the cheapest pipe already exists. Enbridge’s Southern Lights, the larger of the two import lines, added about 10% in 2025 with station work alone. Verified Enbridge planned to take it to 275 kb/d as early as 2013, and said in July that it “can be expanded with future production growth”. Reported That would add about 75 kb/d on an existing right-of-way: the fastest large step. Estimate Reversal is slower. Southern Lights is itself partly a reversed crude line, turned around in 2010 to carry diluent from Minnesota to Edmonton. Reported Doing the same with another Mainline line works only in sequence: space opens when export barrels move to new pipes, which on current plans means Pacific Link; rail carries the gap until then; and the conversion takes months of work and approvals. It is a 2030s option, not a bridge. Interpretation The third route is a return line beside a new export pipe: Northern Gateway was designed to carry 193 kb/d of condensate inland; Pacific Link, on current plans, has none. Reported Producers will sign for a northbound line if its toll beats rail, which at roughly half the cost it should, once enough of them commit together, and a return line is cheapest to add while Pacific Link is still being designed. No single option is enough; the gap will be closed by several at once, or not at all. Interpretation

Every route also has a headwind (Table 4). Pipe faces the fewest on cost and the most on contracts and permits. If none is built in time, the market rations diluent by price: rail sets the cost, growth slows, and that is the outcome that cuts royalties and leaves Pacific Link part-full. Interpretation

Table 4What could slow each fix
RouteWhat could slow itWhat would clear it
Expand Southern LightsShippers must sign before growth is certain; possible US cross-border approvalsThe producers’ growth agreement, then an open season
New or reversed line north10- to 20-year contracts; permits on both sides of the border; a reversal waits for Pacific Link to free Mainline spaceProducer contracts; Pacific Link in service
Return line with Pacific LinkAdds cost to a C$35–44 billion project; needs a coastal import terminalA decision before the design is fixed
RailAbout twice the cost of pipe; railcars, unloading slots and labour disruptionsPrice: rail is the market’s default
Diluent recovery and splittersRail for the heavier crude; buyers for jet fuel and heavy cuts; small scaleGibson’s optional second phase; a refiner or airport offtaker
More LNG demand20-year buyers for plants starting in the 2030s, when the delayed supply wave lands; feed-gas pipe; emissions rulesBuyers paying for a short, chokepoint-free route
Insight Stream interpretation, drawing on the sources for Sections V and VI.

For Pipes, Prices and Governments

The squeeze is measured in billions. Figure 8 puts the main numbers side by side. Estimate

Fig. 8The diluent squeeze in dollars
Annual diluent costs at 2025 prices, and the capital at stake
What it shows → The diluent squeeze in dollars: what Alberta pays for diluent each year, what that bill becomes as the pipes fill, and the capital riding on the outcome.
How to read it → Left: annual flows at 2025 prices, Insight Stream estimates. Right: one-off capital, as reported. Colour: red for costs that grow as the pipes fill, gold for price effects, navy for public money.
Why it matters → A diluent shortfall of 100 kb/d strands about C$6 billion a year of bitumen. Even rail, the costliest bridge, adds only $0.2–0.5 billion a year for the same 100 kb/d.
EVERY YEAR, AT 2025 PRICESCAPITAL AT STAKEC$11.5bnDiluent bought from outside Alberta357 kb/d todayC$18bnWith every oil export pipe full~555 kb/d, central caseC$10bnDiluent to fill Pacific Link alone300 kb/d at a 30% blend~C$6bnBitumen value strandedper 100 kb/d of diluent that does not arriveC$0.4bnCost to producers per US$1/bblwider condensate spread; C$0.1bn of royalties$0.2–0.5bnExtra cost of rail over pipeper 100 kb/d, at the 2015 cost gapC$34bnTrans Mountain, built by Ottawaalready carrying diluted bitumenC$35–44bnPacific Link, government-ledno diluent return line in current plansC$33bnLNG Canada’s second phaseclosing the gap takes two more that sizeC$1.3bnAlberta’s Keystone XL lossthe province has underwritten pipes beforeInsight Stream estimates at 2025 averages: condensate US$63.35/bbl, WCS US$53.70/bbl, US$1 = C$1.40. Capital figures as reported.
Insight Stream estimates from AER ST98-2026 volumes and Alberta Bitumen Valuation Methodology prices (2025 averages: condensate US$63.35/bbl, WCS US$53.70/bbl; US$1 = C$1.40). Rail cost gap: CAPP via the NEB (2015). Capital: Trans Mountain and Pacific Link as reported in the press; LNG Canada phase 2, Office of the Prime Minister (29 Sep 2026); Keystone XL, National Observer (9 Jun 2021).

Pipes are built on contracts but paid for by barrels. Trans Mountain and Prairie Connector are mostly contracted for 15 to 20 years; Enbridge’s Mainline has no contracts, so shippers nominate month to month (Table 5). Verified When a contracted pipe opens before production grows to fill it, its shippers will move barrels off the Mainline. Interpretation Under the settlement that runs to the end of 2028, each 50 kb/d lost raises the Mainline toll by US$0.035 a barrel for those that stay. Verified Only once Pacific Link flows could a Mainline line be spared and reversed to bring diluent in. Interpretation

Table 5Who carries the risk if new pipes run part-full
PipelineHow its capacity is soldWho carries the volume risk
Enbridge MainlineMonth-to-month nominations; contracting rejected by the CER in 2021Remaining shippers, through volume-adjusted tolls (settlement to end-2028)
Trans Mountain (federal Crown corporation)About 80% under 15- and 20-year firm contractsContract shippers; the Crown as owner for the rest
Prairie Connector465 kb/d of 20-year commitmentsContract shippers and South Bow
Pacific Link (government-led)Open season expected in spring 2027Shippers who sign; governments for the rest
Sources: Kinder Morgan via World Pipelines (Trans Mountain contracts, 2017); South Bow Q2 2026 call; CER decision on Mainline contracting (Nov 2021); Enbridge, Mainline tolling agreement (May 2023) and CER approval (Mar 2024); The Hub (Pacific Link).

Governments carry the risk twice, and with them taxpayers. First through royalties: Alberta calculates oil sands royalties after diluent costs, and projects deducted C$29.7 billion of them in 2025. Verified Condensate cost about US$9.65 a barrel more than WCS in 2025 on Alberta’s reference prices; on our estimate that cost producers about C$3.8 billion and cut royalties by about C$0.9 billion, and every US$1 it widens costs producers about C$0.4 billion a year and Albertans about C$0.1 billion. Estimate Second, as owners: Ottawa spent about C$34 billion on Trans Mountain, and Pacific Link, at C$35–44 billion, is government-led. Reported If diluent limits growth, Pacific Link fills last, and every 100 kb/d of diluent that does not arrive strands about 230 kb/d of bitumen worth about C$6 billion a year. Estimate That is public money riding on a diluent plan that does not yet exist; our next article puts a full number on the bill. Interpretation

The shortage reaches well beyond the pipes. Table 6 sets out the direction for each market it touches.

Table 6What a diluent shortage means across the barrel
MarketDirectionWhy
Condensate at Edmonton▲ Premium risesThe marginal barrel comes by rail; the posted price already runs at 95–98% of WTI
Liquids-rich drilling, Montney and Duvernay▲ Keeps growingCondensate pays for the well; the gas still needs a buyer (Article 50)
AECO gas▼ Weaker until LNG demand arrivesLiquids drilling brings gas with it (Article 50); oil sands fuel demand is the offset
Propane and butane at Edmonton▼ Weaker pricesMore supply: liquids-rich wells bring LPG too, and butane blending is capped, so more must be exported
Heavy crude and bitumen netbacks▼ LowerEach barrel carries costlier diluent; royalties are paid after diluent costs
Synthetic crude▲ Worth moreUpgraded barrels need no diluent and can stand in for some
US Gulf Coast condensate and natural gasoline▲ More demand from CanadaMore barrels move north, mostly by rail
Rail and terminals▲ New inbound flowAbout 1.7 unit trains a day for every 100 kb/d
Jet fuel in Western Canada▲ More supply if splitters are builtA third of a split barrel is jet-range fuel
Insight Stream interpretation. Direction is relative to a world with enough diluent; magnitudes are for Part 2.

Who Holds the Cards

The squeeze will not land evenly: a tighter diluent market costs some players, pays others, and only some can act on it (Figure 9). Interpretation

Fig. 9Who gains, who pays, and who can act
Main players in Alberta’s diluent market, by exposure and by the levers they control
What it shows → Who gains and who pays as diluent tightens, and who can act on it.
How to read it → Left to right: from those who pay more to those who gain. Bottom to top: from those with few levers to those who own a pipe, a supply or the contracts. Names are examples, not a ranking.
Why it matters → The fix needs a deal between the top two boxes: the companies that own pipe and supply, and the buyers whose contracts make it bankable.
OWN THE FIX OR THE SUPPLYEnbridgeSouthern Lights and Norlite; says both can expandShell (with ARC)LNG Canada operator; ~107 kb/d oil and condensateGibsonHardisty diluent recovery, +50 kb/d from 2028KeyeraEdmonton condensate hub and diluent rail terminalCPKC and CNRail from the US Gulf CoastPAID BY THE SQUEEZEMontney and Duvernay producerse.g. Ovintiv, buying condensate-rich acreageUS Gulf Coast sellersCondensate and natural gasoline, by railUpgradersSuncor, Canadian Natural: synthetic crude needs nonePAY, BUT HOLD THE CONTRACTSSAGD producerse.g. Cenovus (with MEG), ConocoPhillips, StrathconaTheir leverLong contracts make a diluent pipe bankableTheir exposureAll planned growth is bitumen needing diluentCARRY THE RISKAlberta~C$0.1bn of royalties per US$1/bbl of spreadOttawaC$34bn in Trans Mountain; Pacific Link nextMainline shippersTolls rise if barrels move to contracted pipes← Pays more as diluent tightensGains as diluent tightens →↑ Controls a lever: pipe, supply or contracts↓ Few levers of its own
Insight Stream interpretation, based on company filings and reports cited in the text. Names are examples, not a ranking or an investment view.

Enbridge holds the fastest pipe fix. It owns Southern Lights and Norlite, which carries diluent to the oil sands. Southern Lights’ original 15-year contracts expired in June 2025, and its revenue fell afterwards. Reported A growth wave lets Enbridge re-contract and expand in one step, and a diluent line hedges its Mainline: barrels that leave for contracted export pipes still need diluent that Enbridge can carry in. Interpretation

Shell holds the biggest supply lever. It operates LNG Canada and in September completed its purchase of ARC Resources, which produced about 107,000 b/d of crude oil and condensate in 2025 with 1.3 Bcf/d of gas, about 80 barrels per MMcf. Reported That is equivalent to almost a fifth of Canada’s roughly 570,000 b/d of condensate, on Rystad’s estimate. Estimate How fast Shell develops liquids-rich Montney gas for LNG Canada will move Western Canada’s condensate supply more than any other single decision. Interpretation

Condensate is why the Montney is not for sale cheaply. Ovintiv bought Paramount’s Montney assets for about US$2.3 billion in early 2025 and agreed in November 2025 to buy NuVista for US$2.7 billion; Shell paid US$16.4 billion, including debt, for ARC. Reported Buyers are paying for liquids. A tighter diluent market raises the value of every condensate barrel whatever gas fetches, which is why liquids-rich producers keep drilling and keep their acreage. Interpretation

The companies closest to the problem are already moving. Gibson plans to add 50,000 b/d to its Hardisty diluent recovery unit from 2028, with an option for 50,000 more, within up to C$1 billion of growth capital over five years. Reported Keyera expanded Edmonton’s Alberta Diluent Terminal in 2008 to take more than 60,000 b/d of condensate from unit trains; CPKC links Hardisty and Port Arthur, Texas, on a single railway, and CN reaches the Gulf through its US network. Reported Rail and terminal owners gain most in the bridge years. Interpretation

The buyers pay, but hold the contracts. Much of the planned growth comes from SAGD producers such as Cenovus, which added about 110,000 b/d with its C$8.6 billion purchase of MEG in November 2025, ConocoPhillips and Strathcona. Reported Suncor and Canadian Natural, which upgrade much of their output, are partly hedged: synthetic crude needs no diluent. Interpretation The buyers’ 10- to 20-year contracts are what make a new diluent pipe bankable, and whoever signs first sets the terms. Interpretation

How it lines up. Governments have the fewest quick levers, but Alberta has underwritten pipes before, losing about C$1.3 billion when Keystone XL was cancelled in 2021. Reported The interests point to one deal: SAGD producers sign long contracts, Enbridge expands Southern Lights or builds beside it, and Alberta, whose royalties fall as diluent costs rise, has reason to underwrite part of the bridge. Shell’s Montney pace and the next LNG decisions set its size. The producers’ growth agreement with Ottawa and Alberta is the natural place to strike it, yet nothing reported so far puts diluent on its agenda. Interpretation


What We Expect, and What Would Prove Us Wrong

These are our calls, and what would prove us wrong.

Table 7Our calls
Our callByWhat would prove us wrong
Enbridge moves first: an open season to expand Southern Lights, or to build a diluent line beside it, once producers sign their growth agreement2027No open season by the end of 2027, or the growth agreement stalls
Rail carries the 2029 peak, and Edmonton condensate moves from 95–98% of WTI towards parity or a premium2028–29Growth on the AER’s slower path, or new import pipe before 2029
No Mainline line is reversed for diluent before Pacific Link flowsEarly 2030sA reversal application before Pacific Link is sanctioned
Pacific Link’s open season carries no diluent return line unless governments require oneSpring 2027Return-line or import terms in the open season
No LNG plant beyond Ksi Lisims is sanctioned before 2028, so the roughly 200 kb/d left after Ksi Lisims stays open into the 2030s2028Another LNG Canada-sized decision by 2028; each covers about half the remaining gap on the AER’s yield
Diluent recovery grows first: Gibson sanctions the optional second 50 kb/d phase at HardistyBy 2028The option lapses, or rail capacity for the stripped bitumen is not secured
Insight Stream interpretation. We will report back against these calls.

Three dates will test the balance itself: Ksi Lisims’s investment decision, targeted by the end of 2026 (without it the gap is 90–120 kb/d larger); Prairie Connector’s, targeted for mid-2027; and the AER’s next outlook, due mid-2027, which will show whether the regulator counts the new pipes.

Fig. 10Decide now, bridge to 2030, build for the 2030s
Who should do what, and when
What it shows → What each group should do now, in the bridge years and for the 2030s.
How to read it → Rows are the players; columns are the time windows. Read across a row for one player’s path.
Why it matters → The decisions that matter most, contracts and a return line, have to be made before the pipes fill.
NOW · 2026–27BRIDGE · 2028–30BUILD · 2030sGovernmentsOttawa and AlbertaCount diluent in everypipeline decisionBack diluent recoveryand rail unloadingA return line or importroute with Pacific LinkOil sands producersthe diluent buyersContract diluent beforethe pipes fillRail and recovery tocover 2029 and beyondLong contracts that makea northbound line bankableGas producersand midstreamKeep liquids-richdrilling fundedCondensate handling andfractionation capacityFeed LNG fromliquids-rich wellsPipeline ownersand their shippersOpen season to expandimport capacitySettle Mainline tollsbeyond 2028Design Pacific Linkas a two-way pipeInsight Stream recommendations, based on the balance, forecast and options above.
Insight Stream recommendations.

Canada is debating how to get more oil out of Alberta; the diluent that has to come in is barely part of that debate. Every oil export pipe is also a diluent import requirement, and the gas that could supply the diluent arrives later and only if it is rich. Planning the route in now is cheaper than discovering the shortage when the pipes fill.

IS

Insight Stream — LNG/LPG & Waterborne Trade Analysis

Condensate and diluent balances, LNG contract economics, LPG benchmarks and NGL export infrastructure. Independent market intelligence, built only on sources you can check.

Read More →