In January 2026, Pakistan — a country the LNG industry has spent a decade describing as one of its great future growth markets — had a problem nobody's demand model predicted. It had too much LNG. Terminal utilization had fallen below minimum dispatch levels. Solar had eaten the power demand the cargoes were contracted to serve. The country was exploring how to resell gas it was legally obliged to buy.

Hold that image, because it is the single most important data point in the global LNG demand debate — and it arrived two weeks before Shell published an Outlook projecting demand growth of 65 percent by 2050, to nearly 700 million tonnes a year, with South and Southeast Asia absorbing roughly 40 percent of global imports. Verified

Article 25 in this series argued that the coming supply glut clears on cash cost, not breakeven. This article asks the question that thesis leaves hanging: what if a large part of the demand that is supposed to absorb the wave cannot physically take delivery of it? Not because the price is wrong — but because the terminal was cancelled, the pipeline was never laid, the power plant never signed its offtake, and the turbine is on a seven-year waitlist. The industry models demand as a price curve. On the ground, demand is a construction project. And in the region carrying the forecast, the construction project is failing two times out of three.


2–3× Ratio of shelved/cancelled to
built LNG import capacity in
Southern Asia, past decade
1 of 22 GW Vietnam's planned LNG power
fleet with a signed power
purchase agreement
$5.6B Pakistan's liability on excess
contracted LNG cargoes
through 2032

700 Million Tonnes, Resting on the Hardest Buyers to Build For

Shell's LNG Outlook 2026 — the industry's most-watched demand forecast, now in its tenth year — projects global demand rising from 422 million tonnes traded in 2025 to nearly 700 million tonnes by 2050. It calls for roughly 200 million tonnes a year of additional liquefaction beyond what is already under construction. And it locates the growth squarely in South and Southeast Asia, which it forecasts will account for around 40 percent of global LNG imports by 2050 as those economies switch from coal. Verified

But the Outlook contains its own caveat, easy to miss: the ability to benefit from new supply "will depend on the availability of infrastructure in importing countries, including regasification capacity and pipeline connectivity, especially in South and Southeast Asia." IEEFA, reviewing the Outlook, went further — calling the rationale for rapid demand growth "increasingly fragile," and noting a two-decade pattern the forecast asks us to ignore: outside China, incremental gas demand has overwhelmingly been met by domestic production, while demand has tended to fall in countries that must import at scale. Verified

Both sides argue about price. Bulls say cheap post-glut LNG unlocks Asia; bears say even cheap LNG loses to domestic gas at $1–5/MMBtu and to renewables. This article argues they are both looking past the binding constraint. Before price can matter at all, a cargo must pass through four physical and commercial gates. In the region carrying the forecast, every gate is jammed.


What Actually Stands Between a Cargo and a Kilowatt-Hour

The Four Gates: Where Forecast Demand Dies Between Ship and Socket
Every tonne of "future Asian demand" must clear all four. Failure at any one gate strands the entire chain — and the cargo.
GATE 1 Import terminal S. Asia cancels 2–3× more capacity than it builds (GEM) GATE 2 Pipeline network Bangladesh: only 8% of 2,695 km planned is under construction GATE 3 PPA + turbine Vietnam: 1 PPA signed of 22 GW planned · 7-yr turbine backlogs GATE 4 Solvent buyer Pakistan: $5.6B excess cargo liability · demand destroyed at $15, not bid A price signal can travel this chain in a day. Concrete, steel, and credit take a decade — when they arrive at all.
Framework: Insight Stream · Data: Global Energy Monitor (Mar 2026) · IEEFA · Kpler  |  Data verified  Framework: Interpretation

Gate one: the terminal itself. Global Energy Monitor's March 2026 briefing quantifies what the industry prefers to describe anecdotally: over the past decade, Southern Asia's LNG importers have shelved or cancelled two to three times as much planned import capacity as they have brought online. There is $107 billion of potential terminal and pipeline investment "in development" across the region — but in aggregate, GEM finds there is currently not enough import capacity to realize the demand growth the IEA and others forecast. Even the supposedly fast, cheap fix — floating regas — has stumbled: FSRU projects have been cancelled in Bangladesh and the Philippines and repeatedly delayed in Vietnam, while Bangladesh's existing floating terminals have suffered months-long, cyclone- and mooring-related outages. Verified

The Cancellation Ratio: Planned LNG Import Capacity — Built vs Shelved/Cancelled
Past decade, normalised per unit built · Southern Asia kills 2–3 projects for every one it completes; Europe's ratio is roughly even
0 1× 2× 3× Southern Asia — built 1× Southern Asia — shelved/cancelled 2–3× Europe — built 1× Europe — shelved/cancelled ~1×
Source: Global Energy Monitor, "Southern Asia's gas plans may be overblown" (Mar 2026)  |  Verified

Gate two: the pipeline behind the terminal. A terminal without takeaway is a very expensive pier. In Bangladesh, pipeline connectivity is a named bottleneck: of 2,695 kilometres of pipeline in development, just 8 percent is under construction. India's own history makes the same point — its terminal buildout ran ahead of its pipeline network for years, which is exactly why Kochi and Dabhol sat chronically underutilised while Dahej ran above nameplate. Verified

Gate three: the power plant that never reaches FID. Vietnam's national power plan envisions more than 22 GW of LNG-fired capacity by 2030. As of IEEFA's latest count, exactly one plant has finalised a power purchase agreement. In the Philippines, one greenfield plant has secured offtake. Bangladesh's own state gas company projects no power-sector gas demand growth through 2028. Pakistan abandoned new LNG-fired power plants entirely in 2023 on cost. And now a global constraint compounds the local ones: gas turbine order backlogs stretch to seven years, capital costs have risen from under $1,000/kW to $2,200–2,500/kW since 2022, and two-thirds of gas power projects in development worldwide have not even announced a turbine manufacturer. Verified

Gate four: a buyer who can pay when it matters. The 2026 Hormuz crisis ran the experiment in real time. When spot hit $15/MMBtu, South Asia's response was not aggressive bidding — it was demand destruction of roughly 4 million tonnes: load shedding in Pakistan, industrial gas cuts in Bangladesh, 5–10 percent industrial demand reductions in India. Buyers with one to two weeks of storage cover and strict affordability ceilings do not compete for cargoes; they exit. For the second time in four years — 2022, then 2026 — the marginal Asian buyer proved to be the first casualty of a tight market, not the growth engine of a loose one. Verified

The Affordability Gap: What LNG Must Beat (US$/MMBtu)
Domestic and pipeline gas costs across Asia vs wholesale prices in LNG-dependent markets · before adding regas, pipeline tolls, and distribution
$0 $4 $8 $12 $16 Vietnam — associated gas $1.00–1.25 Pakistan — wellhead $3.50–4.00 Thailand — offshore ~$5 Indonesia — piped to Sing./Malaysia ~$6 China — pipeline imports $8–9 Japan/Korea — LNG-dependent wholesale $11–15
Source: IEEFA Data Dive — drivers, barriers and costs of Asia's gas and LNG demand  |  Verified
The Proof Case Nobody Modelled · Pakistan · January 2026 Surplus

Weeks before the Hormuz crisis, Pakistan — the archetypal "future growth market" — held an LNG surplus: terminal utilization below minimum dispatch, solar eating the power demand the cargoes were contracted to serve, and an estimated 177 excess contracted cargoes through 2032 carrying a liability of $5.6 billion or more. Demand did not fail to appear because LNG was too expensive that month. It failed to appear because the buyer's power system had already been built around something else.

IEEFA · The Diplomat (Mar 2026) · Verified

A World Drowning in Regas Capacity — In All the Wrong Places

Here is the twist that makes this a structural problem rather than a timing one. The world does not lack regasification capacity. The IGU's 2026 report counts 1,113.5 Mtpa of it — running at 39.2 percent average utilization. On paper, the world could nearly triple its LNG imports through existing terminals. Verified

But capacity does not travel. Japan holds 219 Mtpa of it at 31 percent utilization — a mature market in structural decline. South Korea holds 143.5 Mtpa at 34 percent. Europe built a crisis surplus after 2022 that is now chronically underused as its gas demand falls. Meanwhile the markets that are supposed to carry the next 280 million tonnes of demand growth hold slivers: Bangladesh 8.3 Mtpa, Pakistan 10 Mtpa, Vietnam 4 Mtpa, the Philippines 8 Mtpa — and, as Section II showed, their pipelines to expand are 2–3 times more likely to die than deliver. The industry's comfort statistic — "plenty of regas capacity globally" — is a map error. The spare capacity sits where demand is shrinking; the missing capacity sits where demand is forecast. Interpretation

Regas Capacity vs. Where the Growth Is Supposed to Happen (Mtpa, end-2025)
Mature, declining markets hold the world's spare import capacity; the forecast growth markets hold slivers of it — and cancel most of what they plan
0 100 200 Mtpa Japan (31% util.) 219 South Korea (34%) 143.5 China (40%) 175 (+116 building) India (47%) 52.5 Pakistan 10 ← the "growth markets" carrying B'desh · Phil. · Vietnam ~20 combined 40% of Shell's 2050 import forecast
Source: IGU 2026 World LNG Report (capacity & utilization) · Shell LNG Outlook 2026 (forecast share)  |  Verified

India Proves the Rule — Demand Grows Where the Chain Exists

◆ Where the Bulls Are Right

India is the strongest rebuttal to this thesis, and it deserves a fair hearing. The IEA projects Indian gas consumption rising nearly 60 percent by 2030, with LNG imports roughly doubling to 64 bcm/yr; nearly 40 bcm/yr of new regas is under construction or planned; city gas distribution networks now cover essentially the entire mainland; and terminal utilization is improving. If India delivers, it alone absorbs a meaningful slice of the coming supply wave. Verified

But look at why India works, because it proves the framework rather than breaking it. India's demand is growing precisely because it spent two decades building all four gates: eight operating terminals, a 25,000-kilometre operational pipeline network with 33,500 authorised, 307 city-gas geographical areas covering the whole country, anchor demand in CGD and fertiliser with administered pricing, and creditworthy state-backed offtakers. Even so, the binding constraint remains infrastructure, not price: Indian terminals ran near 50 percent utilization in 2024 — not for lack of cargoes but because of pipeline gaps and downstream offtake limits, and India's own regulator has flagged terminal tariffs as a barrier to adoption. When ORF modelled a sustained $6 LNG world, the demand uplift was 5–10 percent above baseline — meaningful, but "incremental rather than transformative," bounded explicitly by pipelines and last-mile capacity. Verified

The Counter-Case: India's Gas Demand and LNG Imports, 2023 vs 2030 (bcm/yr)
Growth of ~60% in demand and a doubling of LNG imports — delivered by two decades of infrastructure, not by price alone. Even here, terminals ran near 50% utilization on pipeline and offtake limits.
0 50 100 bcm 65 2023 demand 103 2030 demand +~60% 36 2024 LNG imports 64 2030 LNG imports ~2× — 4th-largest importer
Source: IEA India Gas Market Report (Outlook to 2030) · PNGRB  |  Verified

That is the whole argument in one country: even in Asia's best-built gas market, cheap LNG buys you single-digit upside against the infrastructure baseline. Price is the accelerator. Infrastructure is the engine. The forecast assumes the accelerator can substitute for the engine. Interpretation

2–3× Southern Asia's shelved/cancelled import capacity vs. built, past decade — vs ~1:1 in Europe

Global Energy Monitor, Mar 2026 · Verified

$1–5 Domestic gas cost (US$/MMBtu) in Vietnam, Pakistan, Thailand — the price LNG must beat, before regas and pipelines

IEEFA Data Dive · Verified

7 yrs Gas turbine delivery backlogs; two-thirds of projects in development have no turbine maker announced

IEEFA · GEM/Data Desk · Verified

~4 Mt South Asian demand destroyed at sustained $15/MMBtu during the 2026 crisis — exit, not bidding

Kpler assessment, Mar 2026 · Verified


The Glut's Demand-Side Mirror — and Who Actually Wins

Connect this to Article 25 and the picture completes. The supply side of the 2027–2030 glut clears at the US Gulf cash-cost floor. The demand side, this analysis suggests, cannot expand fast enough to absorb the wave even at that floor — because the marginal buyer's constraint is not price but concrete, steel, contracts, and credit. Two consequences follow, and neither is the one the consensus draws.

First: the glut runs deeper and longer than price models imply. Most oversupply models assume low prices summon new demand within quarters. But if the demand response requires terminals that fail two times in three, pipelines at 8 percent construction, PPAs nobody signs, and turbines on seven-year waitlists, the demand response is measured in half-decades. Cheap LNG cannot buy what has not been built. Interpretation

Second: the fight concentrates on the demand that already exists. If the growth markets cannot take delivery, the wave's absorption falls to the buyers whose four gates are already open — China's expanding coastal system (116 Mtpa of regas under construction, more than half the global total), India's functioning chain, and the mature Northeast Asian markets replacing legacy contracts. That is a brutal, concentrated battleground of creditworthy buyers with real infrastructure — and it is fought on delivered cost and reliability, which is exactly where short-haul Pacific suppliers with oil-linked term books are structurally strongest. The last-mile problem does not soften Article 25's conclusion. It sharpens it. Interpretation

The Commercial Implication — Sell to Infrastructure, Not to Forecasts

For suppliers and portfolio players, the practical lesson is that the next decade's reliable demand sits behind gates that are already open. Underwriting offtake against a forecast of South and Southeast Asian growth means underwriting terminal completion, pipeline construction, PPA execution, turbine delivery, and sovereign credit — five risks the seller cannot control. The Japanese trading-house model of the 1970s–90s — and China's integrated buildout today — points the other way: the demand that materialises is the demand someone financed end-to-end. Suppliers who want the 700-million-tonne world will have to help build it, gate by gate, or plan for a smaller one.

⚠ Where This Thesis Can Break

Three honest risks. One: a deep, multi-year price trough could change political calculus — sustained sub-$6 LNG makes terminals easier to finance and PPAs easier to sign, and India's own data shows real (if bounded) elasticity. Two: FSRU economics could improve as the post-2022 charter squeeze unwinds, lowering gate-one costs. Three: China is the wildcard that can absorb far more than modelled — its 116 Mtpa regas buildout is the single largest demand-side de-risking event in the market, and it is real, funded, and under construction. If China over-delivers, the wave finds a home regardless of what Dhaka and Hanoi build. The thesis bounds the forecast; it does not doom it.


Demand Is a Construction Project

The LNG industry has spent a decade telling itself a price story about Asian demand: when the wave of supply arrives and prices fall, the buyers will come. Shell's 700-million-tonne world is the fullest expression of that story. But the record from the very region carrying the forecast tells a construction story instead — of terminals cancelled at two and three to one, pipelines that exist mostly on paper, a 22-gigawatt power plan with one signed offtake, and a flagship growth market that greeted 2026 with a surplus of gas it could neither burn nor afford.

Price signals travel at the speed of a trading screen. Concrete travels at the speed of permits, credit committees, monsoons, and turbine queues. The gap between those two speeds is the last mile problem — and it, not the supply wave, may prove to be the defining constraint of the LNG market's next decade. The cargoes are coming. The question the industry has not answered is who, physically, will be able to take them.

Demand is not a price curve. It is a construction project. And right now, the project is behind schedule.

IS

Insight Stream — LNG/LPG & Waterborne Trade Analysis

26 articles covering LNG demand and oversupply economics, China PDH, VLGC freight markets, LPG/NGL infrastructure, energy security, and global waterborne trade. Independent market intelligence from a practitioner perspective.

Read More →