Picture a shipyard and you picture heat. Sparks arcing off a hull, plate the size of a tennis court being cut and bent, something glowing in the middle distance. It looks like exactly the kind of place that should dread an energy shock. When prices spike, surely a shipyard is near the front of the queue to feel it.
It is nearly the last. A modern yard is closer to an enormous, very heavy assembly hall: it runs mostly on electricity, it can pause and restart without wrecking anything, and its own energy is a low single-digit share of the cost of a finished hull. In the first half of 2026 Korea's big shipbuilders watched their power tariffs climb and posted record profits anyway.
So the interesting question is not how much energy a shipyard uses. It is how much energy it buys without noticing, and who it sells the results to.
The answer, in both directions, is that a shipyard's real energy bill is one it never pays. It arrives already spent, baked into the steel plate it buys at a price agreed years earlier, and it arrives again as the fuel and carbon rules that decide which ships anyone orders at all. This is the whole of the risk, and almost none of it would show up on an energy audit.
Where the energy actually sits
Start inside the fence, because that is where the intuition goes wrong. A yard's own processes span cutting, bending, welding, blasting, painting and coating, yet building a hull accounts for only about 4 per cent of that ship's lifetime carbon. The work is overwhelmingly electric. A stubborn residue still burns oxygen and acetylene gas to cut and straighten steel with a flame, something no modern kit has fully replaced.
Welding shows the split precisely. Oxyfuel welding, which joins steel with a gas flame, takes about 47 per cent of welding energy; arc welding, which uses electricity, takes 39 per cent; and resistance welding takes the rest. So a large slice of the yard's joining and cutting work runs on bottled gas, not on electricity. That shields part of a yard from a blackout, but exposes it to the industrial-gas supply chain instead. Either way, it is a modest bill. A yard can absorb a sharp rise in its own power cost far more easily than a rise upstream.
Note. Most of a yard runs on electricity, but a large share of joining and cutting energy is still delivered as bottled gas rather than electrons.
The steel is the bill, and the contract is the trap
A hull is 75 to 85 per cent steel by weight, and steel plate is priced off iron ore, coking coal, freight and power. So when coal and electricity prices move, they reach the shipyard through the mill that rolled its plate, not through the meter on its own wall. Marine-grade plate, the certified kind a classification society signs off, such as LR Grade A, has been running at roughly 690 to 760 US dollars a tonne. Competition among Asian mills has held the increases down so far. The dependency is also geographic, and slightly awkward: Korean yards, which build the most valuable ships in the world, lean partly on Chinese steel, a trade worth about 5.63 billion US dollars in 2024.
Here is where an ordinary cost problem becomes a structural one. Most large newbuild contracts are fixed-price and multi-year: a yard agrees a number today for a ship it hands over in two or three years, and if steel or power costs rise in the meantime, the builder absorbs the difference. There is no surcharge and no repricing. A commodity trader adjusts prices weekly; a shipbuilder has already sold the thing. Cost escalation is immediate, and the ability to reprice is deferred to the next contracting cycle.
What is holding the line is an unusually deep buffer. South Korea's three major builders posted a combined second-quarter 2026 operating profit of 2.71 trillion won, a record, on order backlogs near 210 trillion won, roughly three years of secured work. HD Korea Shipbuilding alone reported second-quarter operating profit up 72.5 per cent year on year. That backlog is the whole risk picture in one figure. Three years of work, signed before any 2026 shock, lets a builder absorb serious cost pressure on hulls it has already sold. But the buffer is cyclical, and it is thickest in exactly the producers whose costs are most energy-sensitive.
Three countries, three grids
Shipbuilding is among the most geographically concentrated heavy industries on earth. China, South Korea and Japan build almost all of the world's tonnage, so the industry effectively rents its energy resilience from three national grids. The concentration is not gentle: in the first quarter of 2026 Chinese yards took 84.9 per cent of new orders by deadweight tonnage, the standard measure of the cargo a ship can carry, leaving South Korea 12.8 per cent and Japan 1.4 per cent, the lowest Japanese share in decades. There is no meaningful spare capacity anywhere else, so an energy event that idles Korean or Chinese yards cannot be routed around.
The two grids that carry the load pose opposite risks. China's is enormous and cheap, but it runs on coal and is periodically rationed. In 2024 the grid switched off 5 to 12 per cent of the country's wind and solar power rather than let it overload the system, and the coastal belt where the yards cluster carries a tail risk of ordered power cuts in a crunch. Korea's grid is reliable but expensive, and getting more so. Almost all its industrial power comes from a single state utility, KEPCO, which lost around 43 trillion won between 2021 and 2023 and has passed the cost on. It lifted industrial tariffs by an average of 9.7 per cent in October 2024, and the heavy-user band the shipyards sit in absorbed 10.2 per cent. So the two face opposite threats: a Chinese yard, a small chance of being switched off outright; a Korean yard, a near-certain, gradual ratchet on power prices, even with the order book full.
Note. Three countries build almost every ship, and there is no spare capacity elsewhere, which is why a regional power event becomes a global one.
The oil shock that cuts both ways
Oil has run a war premium all year. A blockade of the Strait of Hormuz, the sea lane through which roughly a fifth of the world's oil moves, pushed Brent crude from about 71 dollars a barrel to a peak near 94 in early 2026, and it still traded near 89 in August. For most industries that is simply bad news. For shipbuilding it is genuinely mixed. A crude spike does not by itself push a yard into loss, because direct energy is a small share of its cost base beside steel and labour. The sharper sensitivity sits on the demand side.
What matters to a yard is not the oil price but its gap to the price of gas, and that gap has widened in the yards' favour. Gas is cheap. The main American gas benchmark sat near 2.77 dollars per million British thermal units in August, a fraction of what the same energy costs as oil. That cheapness does two things. It moves gas: exporters ship more of it to buyers who lack it. And it moves owners: they order vessels that run on gas rather than pricier, dirtier oil. Both favour what South Korea builds best, the tankers that carry liquefied natural gas and the dual-fuel ships that can burn either. The same oil spike that raises a Korean yard's costs deepens the order book for its most profitable product.
The real danger is not expensive energy but a demand slump behind it. A deep enough crisis in freight and trade would stall ordering across every vessel type, even as the yards stood ready to build, as happened after the 2008 crash. The timing is cruel. A cost shock arrives in weeks, a demand shock in quarters, and order books this deep mean one landing now would not empty the berths until the late 2020s. By then it is too late to fill the hole.
Note. The war premium is real but not extreme. What the yards fear is not this line, it is a demand slump behind it.
The transition pays the yards it does not squeeze
Now the part that confounds intuition. Decarbonisation, which everyone assumes will punish an energy-hungry industry, is for the high end an engine of demand. The European Union now makes ships pay for the carbon they emit on European voyages, through a market it calls the Emissions Trading System. That charge phased in over three years, from 40 per cent of a ship's emissions in 2024 to 70 per cent in 2025 and 100 per cent in 2026, and it pushes owners toward efficient, alternative-fuel designs. Those happen to be the most profitable ships a yard can build. In the first half of 2026, LNG took 73 of 137 alternative-fuel orders; ammonia, methanol, ethanol and a single hydrogen ship shared the rest. Korea's top three yards hold an estimated 70 per cent of global LNG-carrier production.
So the same conditions that raise a Korean yard's steel and power bill also strengthen the case for the ships it sells. The net effect turns entirely on what a yard builds. A yard turning out standard bulk carriers gets the cost pain and none of the relief; a yard building gas carriers gets both. The catch is that the buyer must pick a fuel and live with it for twenty years. Methanol was the fashionable bet in 2023, and the market quietly walked away from it, stranding both the owner who bought the ship and the yard that tooled to build it. The demand is strong. The exposure is choosing the wrong future to build for.
Note. The order book has bet heavily on one fuel, and the buyer commits for twenty years at the moment of ordering.
The winners are the backlog and the engine makers
If you want to see where the transition is actually paid out, look one step up the chain, at the engines. Hanwha Engine and HD Hyundai Marine Engine booked more than 2 trillion won of orders in the first half of 2026, matching the whole of the previous year, as demand for dual-fuel and alternative-fuel engines surged. Their operating margins widened accordingly. As owners hedge against future fuel and carbon costs, the premium moves to the engine makers only a handful of firms can rival.
That concentration is a double edge. It hands the makers pricing power and margin now, and it makes them a single point of failure later. Two components prove the point. The engines themselves, and the advanced cryogenic containment, the specialist insulation that keeps liquefied gas cold in a ship's tanks, both sit with a short list of suppliers. So a shock reaches the most profitable part of the order book through a very narrow channel. The industry is consolidating, not by merger but by capability. The firms that can build and power low-carbon ships are pulling away; the rest are being pushed toward repair, coastal and defence work, or out of merchant shipbuilding altogether.
Note. The sharpest margin gains sit one step up the chain, in propulsion, where only a handful of firms can supply what the transition demands.
Two dates that will move more money than any barrel
The most consequential events in this story are not energy events. Two of them sit in the autumn of 2026. On 9 November, an American suspension of special port fees on Chinese-built and China-linked ships expires. Those fees, levied under a trade measure called Section 301, would restart on a rising scale: 50 dollars for each ton of a ship's capacity, climbing 30 dollars a year toward 140 by 2028. If the suspension lapses rather than rolls over, a ship built in a Chinese yard can become commercially unattractive to any owner trading into the United States, however well it was built.
The other is a vote at the International Maritime Organization, the United Nations body that regulates shipping. At issue is its Net-Zero Framework, a global standard for ship fuels that would also put a price on their carbon. The vote was deferred once, when 57 states postponed it in October 2025; it is now due in October 2026, and any rules would take effect no earlier than 1 March 2028. Until owners know how LNG, ammonia and methanol will be scored, a share of newbuild decisions simply stays parked. Regulatory uncertainty is doing what a price spike cannot: freezing orders outright. Two administrative decisions, weeks apart, each able to reprice large parts of a global order book, and each, unusually, a risk you can put in a diary.
| Period | Status | Key development |
|---|---|---|
| Aug to Nov 2026 | Elevated, policy-timed | Section 301 fee suspension expires 9 Nov; IMO Net-Zero vote due October; oil elevated after Hormuz |
| Late 2026 to mid-2027 | Margin pressure | EU ETS reaches full phase-in for shipping; oil projected to ease; steel-input sensitivity |
| 2027 to 2029 | Structural re-pricing | Section 301 fee plateau (Apr 2028); earliest IMO entry into force (1 Mar 2028) |
| 2029 to 2036 | Transition-defined | Low-carbon fuel economics and yard decarbonisation set the competitive geography |
Note. About three-quarters of the probability sits on the two mildest paths; the severe tail needs a compound shock, not a single one.
What it comes down to
Shipbuilding in 2026 is commercially strong and structurally exposed, and the two are not in tension. Order books are the fullest in seventeen years, and margins the best in longer than that. That strength is exactly what lets the industry sit downstream of chokepoints it does not control, buy its largest input from an energy-priced market, and sell to customers whose economics are set by that same market.
It will not break in a single energy shock. A welding bay pauses and restarts; a blast furnace freezes. The report's own scenario weighting puts about three-quarters of the probability on the two mildest paths, where energy stress erodes margins and softens orders without breaking the value chain, and leaves a combined 26 per cent tail for the severe outcomes, each of which needs a compound shock rather than a single one.
So the risk here is not a bang. It is that a shipyard's exposure is almost entirely imported, in the price of plate it agreed to years ago and in the fuel economics of customers it cannot control, and that its two sharpest risks are decisions taken by a UN agency and a trade representative. The industry will simply find out, slowly and then all at once, which of its bets were right.
Figures drawn from TheRiskAgent's report Energy Risk for Industries: Shipbuilding (August 2026), which sources each one to a named publisher. Produced with AI research tools and reviewed before release; reference material, not advice.