← Back to Weekend Edition
China's Coal Buildout Has Already Changed the LNG Map
JKM, spot LNG delivered to Japan, closed Friday at $16.07 per MMBtu. TTF, the European gas benchmark, closed at $45.19. The $29 gap between them is wider than it has been at almost any point in the LNG market's history, and the standard explanation does not fully account for it.
The standard explanation is Hormuz. The Strait of Hormuz crisis shut in roughly a fifth of the world's monthly LNG supply earlier this year. That pushed European gas prices sharply higher, diverted Atlantic cargoes west, and left Asian buyers scrambling for alternatives. JKM fell as Asian demand absorbed the shock through inventory drawdown and fuel switching. Eventually Hormuz reopens, the premium fades, and the gap closes. Morgan Stanley's latest Brent cut, to $75 per barrel for Q3 2026, down $15 in two weeks, its second downward revision since the US-Iran memorandum of understanding in June, signals that the oil market at least believes the reopening is underway.
But for LNG, the spread is also doing something structurally different, and the two stories need to be separated.
In 2025, China built 78% of all new coal power generation capacity added anywhere on earth. That number should give pause to anyone modelling Asian LNG demand in 2030. It is not a year's anomaly that reverts when energy security concerns ease. Coal plants have 25-to-40-year operating lives. Every gigawatt commissioned in 2025 under emergency energy-security conditions is a gigawatt that does not run on LNG for the next three decades. The financing is done. The permits are issued. The physical infrastructure is being connected to grids.
The thermal coal import surge now running through Northeast Asian ports confirms this is a current phenomenon, not just a future pipeline. China, Japan and South Korea are all buying thermal coal. Newcastle physical coal at $121 per tonne has made coal-fired generation substantially cheaper than gas at elevated spot LNG prices, and those buyers have been locking in the economic logic with new capacity that burns $121 coal, not $16 JKM.
Shell published its 2026 LNG Outlook last week, projecting a 65% rise in global LNG demand to nearly 700 million tonnes per year by 2050. The company has framed this as resilience in the face of the Hormuz disruption, new North American liquefaction capacity, improved plant performance and the eventual recovery of Strait traffic will, Shell argues, get LNG trade back to 2025 levels by year-end if flows normalise. Long-term, the 65% growth case rests on expanding Asian demand, with China as a foundational buyer.
The 78% coal buildout figure sits in direct tension with that forecast. Shell is projecting growth from a customer actively reducing its structural exposure to LNG purchases. The number to watch is not the 65% headline, it is how much of that 700 Mt depends on Chinese import growth, and what the forecast looks like if China's coal buildout runs for another five years at anything close to its 2025 pace.
The counter-case is not frivolous. Chinese LNG demand has components that coal cannot serve: urban gas distribution networks, industrial feedstocks, coastal and inland cities where coal plants cannot be sited. JKM at $16 reflects compressed spot demand, not zero demand. And if Hormuz reopens fully, the rebalancing of Atlantic cargoes back east could tighten Asian spot prices from the supply side, pushing JKM higher even as Chinese incremental demand growth slows.
The problem with this view is asymmetry. Coal capacity once built creates a floor on coal demand that is very difficult to erode, operating costs are low and sunk costs are irrecoverable. The 2025 coal buildout has structurally raised Asia's coal consumption baseline. When Hormuz normalises and Atlantic LNG cargoes swing east seeking the JKM market, they will find buyers, but at the margin those buyers have 2025-vintage plants burning coal at $121 per tonne. The arb compression, TTF falling toward JKM rather than JKM recovering to meet TTF, is the more plausible direction of travel.
European gas at $45.19 still carries a Hormuz premium. EU storage is running at 49.2%, Germany at 42%, the Netherlands at just 26%, injection is running hard, but Europe has been benefiting from the diversion of Atlantic LNG that would otherwise have moved to Asia. When that diversion reverses, TTF loses some of its supply support and European power prices, currently running at €145 for German Day-Ahead and €149 for Dutch, face compression. The EUA carbon price at €79.45 is consistent with current European power economics, but it cannot price the structural reality that coal capacity in Asia is now growing faster than at any point since China's infrastructure buildout decade of the 2000s.
Morgan Stanley's oil forecast revision, citing weak Chinese crude purchases alongside the Hormuz reopening, fits the same pattern. Chinese energy demand is restructuring. The transition is commodity-specific: oil weakens on surplus and demand softness, LNG faces structural displacement in Asia, coal strengthens on capacity lock-in and import demand. A macro framework that treats fossil fuels as a single direction of travel misses the divergence.
The JKM/TTF spread will close when Hormuz fully reopens and Atlantic volumes rebalance. But the equilibrium will settle closer to $16 than to $45. Shell's 700-million-tonne forecast was built on demand assumptions that 2025's coal buildout has already begun to undercut.
Opinion
2026-07-03 22:30
·
4 min read
Opinion: China's Coal Buildout Has Already Changed the LNG Map
China's Coal Buildout Has Already Changed the LNG Map
Share
More from this Weekend Edition
Big Story
Big Story: The Ceasefire Unwinds Asia's Energy Security Pivot
Opinion
Opinion: Shell's LNG Forecast Is Right in the Long Run, and Wrong Right Now
Opinion
Opinion: Japan's Nuclear Restart Will Move Europe's Winter Gas Price
Opinion
Opinion: OPEC's price problem starts inside its own borders
Got Wrong
What We Got Wrong: What We Got Wrong, Week of June 30 to July 3, 2026
Thematic
The Week Ahead: AI's Inference Turn and the Grid It Will Need