Spend an afternoon reading equipment lists for LNG projects and a pattern emerges. Most of what a liquefaction plant or a carrier is made of can be bought from many suppliers: steel, pumps, valves, switchgear, civil works, accommodation. Ordinary heavy industry, competitively supplied.
Then there are three places where the list of names stops being long.
One: the cold end of a liquefaction train
The refrigeration cycle itself is not a secret. The thermodynamics are in textbooks, and several firms will license you a process. What is scarce is the ability to build the object at the centre of it.
A large coil-wound heat exchanger is a tower containing many kilometres of tubing wound around a core, fabricated to tolerances that hold at minus 162 degrees through decades of thermal cycling. Two firms dominate their manufacture. Air Products states it has shipped more than 115 large coil-wound exchangers over the industry’s history; Linde says it has delivered over a thousand coil-wound units across all sizes. Those are made in a handful of factories.
The alternative construction, brazed aluminium plate-fin blocks, has its own short list, with Chart Industries the best known. Different technology, similar concentration.
Underneath sits a further constraint that gets less attention: the drivers. The gas turbines and compressors that turn the refrigerant circuits come from a small group, and the electrical equipment for an all-electric plant competes for transformers and variable speed drives with every other industrial electrification project on earth. A plant can be permitted, financed and contracted and still wait on a compressor train.
Two: the membrane
An LNG carrier’s cargo touches a metal skin about a millimetre thick, backed by insulation, leaning on the hull for structural support. That membrane has to stay leak-tight across hundreds of cooling and warming cycles while the metal around it shrinks by tens of centimetres.
Most of the modern world fleet uses one of two designs, both licensed by a single French firm. There is a Korean alternative, KC-1, developed by KOGAS with three shipyards precisely because a single licensor for a strategic industry made people uncomfortable, and it has been fitted to a small number of ships. Free-standing tanks, the Moss sphere and IHI’s prismatic SPB, remain available and are rarely chosen for new conventional carriers.
Go one layer down and it narrows again. One of the two membrane systems uses Invar, a nickel-iron alloy that barely moves with temperature, and the supply of that alloy for this purpose is itself concentrated. So is the cryogenic plate for onshore tanks, split between a few mills making 9 per cent nickel steel and, more recently, high-manganese alternatives.
There is no drama in any of this. It is simply what happens when a product needs decades of accumulated process knowledge and has a customer base of a few dozen orders a year.
Three: the yards
Many yards can build a large ship. Very few can build an LNG carrier, and the reason is the second point above: installing a membrane means hundreds of thousands of precise welds in confined spaces, to a standard where a slow leak is a decades-long problem. That capability takes years and licensor qualification to develop, and early ships from a new entrant carry real risk.
The result is that LNG carrier construction sits with a small group of yards in South Korea, China and Japan, with Russia building ice-class tonnage for its own projects because ordering it elsewhere became impossible. Their berths are a fixed resource. When ordering surges, the queue extends by years, and a project needing ships in three years may find the slots gone.
That queue is as real a constraint on LNG supply growth as the plant itself, and it appears in no capacity statistic anywhere.
What follows from this
Two things, one analytical and one about how this site is built.
The analytical point: liquefaction capacity forecasts assume an equipment and shipbuilding supply chain that is not obviously elastic. A wave of final investment decisions does not create exchanger factories or yard berths. When many projects reach construction together, they compete for the same short lists, and the binding constraint moves from money and permits to manufacturing slots. That constraint is invisible if you only count tonnes.
The practical point: this is why a directory of LNG assets is incomplete without a directory of the firms behind them. Our supplier list is deliberately short and hand-checked, with a public source behind every claim, and firms we could not source are recorded as unsourced rather than quietly included. It is a smaller thing than the asset data and a slower thing to maintain, but the asset data alone will not tell you where the chain is thin.
If you want the mechanisms rather than the firms, the explainers on liquefaction trains, containment systems and who builds LNG ships cover each of the three in more depth.