Analysis

LNG Canada’s September 2026 Phase 2 decision pairs two additional liquefaction trains with work on the 670-kilometre Coastal GasLink route. The owner says it will manage pipeline expansion that includes five new compressor stations. Fluor separately announced the plant engineering and construction notice to proceed. Together, the notices describe two connected investment schedules, but they disclose no verified expanded pipeline flow rate.

This makes a useful engineering problem: a liquefaction train accepts a continuous feedgas stream, while a long pipeline delivers gas under pressure and gas-quality constraints. A completed train cannot manufacture the upstream molecules it needs. Conversely, pipeline capacity built earlier than the train may wait for liquefaction and shipping. Coordinating both sides matters more than comparing their announced milestones in isolation.

Evidence and measurement

Pipeline flow depends on inlet and outlet pressure, pipe diameter, length, gas composition, temperature and compressor operating limits. Five stations indicate a planned capacity intervention, but station count cannot be converted directly to an additional billion cubic feet per day. That calculation needs compressor power and curves, allowable pressure, hydraulics, inlet conditions and maintenance margins.

Liquefaction also takes energy and gas treatment. Raw feed volume is not equal to finished LNG energy: contaminants must be removed and some energy powers cooling. Without project-specific conversion and fuel disclosures, a claimed exact feedgas requirement for Phase 2 would be false precision. Procurement should specify delivered gas quality and pressure, not merely headline reserves.

Engineering chain

For an illustrative bottleneck model, express both sides as LNG-equivalent tonnes per year after all conversion losses. Assume a new train pair can make 14 Mt/year and the incremental feedgas system can support 10 Mt/year. The chain can then deliver at most min(14,10)=10 Mt/year before downtime. The 4 Mt/year gap is unavailable nameplate, not proven actual loss.

If hypothetical feedgas capacity rises to 12 Mt/year while trains remain at 14, potential output rises to 12 Mt/year. If the same system then operates at an assumed 85% overall availability, 12×0.85=10.2 Mt/year. The example avoids converting cubic feet to tonnes because LNG composition and process fuel are unspecified; both inputs deliberately use the same LNG-equivalent unit.

Economic mechanism

The economic penalty of mismatch depends on the contract. A buyer may pay reservation charges for unused pipeline service; a terminal may carry interest during commissioning; suppliers may redirect gas. None of these costs is quantified in the two releases. The useful budget comparison is the cost of another marginal unit of deliverable throughput against the net contribution it enables.

As a hypothetical sensitivity, at US$20 per tonne contribution, closing a 2 Mt/year bottleneck would add at most US$40 million per full operating year before capex, taxes and downtime. At US$50 per tonne, it would add US$100 million. These are 2 million tonnes multiplied by assumed unit contributions, not published margins or a business case for a compressor.

Illustrative sensitivity

Storage and berth availability introduce further queues. Even if pipeline gas and trains align, a delayed ship can fill storage and force a temporary rate reduction. Conversely, extra berth capacity alone cannot raise output when inlet gas is scarce. A whole-chain model should include seasonal pipeline performance, train outages, tank inventory and vessel scheduling.

Construction dependencies should be tested with dated commissioning gates: gas receipt, compressor test, gas-treatment acceptance, refrigeration startup, tank readiness and first cargo. First cargo tests the chain once; sustained monthly output tests reliability. Measured production and maintenance history, rather than station count, will show how close the system comes to its design rate.

Uncertainty and decision gates

The evidence cutoff is 6 October 2026. The primary notices do not publish detailed pipeline hydraulics, exact train feed requirements, outage rates or a first full-year output forecast. The calculations above deliberately use hypothetical LNG-equivalent capacities and do not imply actual Coastal GasLink throughput.

For 2026–27 planning, the decision trigger is a jointly tested integrated schedule and firm gas-delivery capability at specification. If either slips, update the cash-flow start date and throughput model. Once measured compressor and train data appear, replace illustrative capacities with the verified operating envelope.