Watch the steel, not the sparks

Energy-transition risk in the global shipbuilding industry.

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, and it is reasonable to assume that when prices spike 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, with a stubborn residue of oxygen-and-acetylene flame for thermal cutting and flame-straightening that no amount of modern kit has fully replaced.

Welding shows the split precisely. Oxyfuel gas welding takes about 47 per cent of welding energy, arc welding 39 per cent, and resistance welding the rest, which means a large slice of the yard's joining and cutting work runs on bottled gas rather than on electrons. That insulates part of a yard from a blackout and exposes it instead to the industrial-gas supply chain. Either way, it is a modest bill. A yard can absorb a sharp rise in its own power cost far more easily than it can absorb what happens 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 classification-society-approved kind such as LR Grade A, has been running at roughly 690 to 760 US dollars a tonne, with Asian mill competition holding 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 already sold. It is also cyclical, and concentrated in exactly the producers whose cost base is 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, 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 anchored on coal and periodically rationed, with wind and solar curtailment of 5 to 12 per cent recorded in 2024 and a coastal yard belt that carries a tail risk of administrative power cuts during a crunch. Korea's 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 across 2021 to 2023 and has passed the cost to industry, lifting industrial tariffs an average 9.7 per cent in October 2024 with the heavy-user category the shipyards sit in absorbing 10.2 per cent. So a Chinese yard faces a lower-probability, higher-severity risk of being switched off, while a Korean yard faces 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, through which roughly a fifth of the world's oil moves, pushed Brent 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, because a crude spike does not by itself push a yard into loss: direct energy is a small share of its cost base beside steel and labour, so the sharper sensitivity sits on the demand side.

Dear oil paired with cheap gas, and United States Henry Hub sat near 2.77 dollars per million British thermal units in mid-August, is precisely the signal that favours the ships Korean yards are built to make. The genuine danger is not expensive energy but a demand slump behind it: a prolonged, disorderly crisis that chilled the wider freight and trade economy would stall ordering across every vessel type even as yards remained able to build, the pattern that followed the 2008 credit freeze. Cost shocks arrive in weeks; a demand shock arrives in quarters, and the deep order books mean one landing now would not empty berths until the late 2020s. By the time the hole is visible, it is too late to fill it.

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 Emissions Trading System reached full phase-in for shipping in 2026, the end of a ramp from 40 per cent of costs in 2024 to 70 per cent in 2025 and 100 per cent now, and that pushes owners toward efficient and alternative-fuel designs, which are the most profitable ships a yard can build. In the first half of 2026 LNG took 73 of 137 alternative-fuel orders, with ammonia, methanol, ethanol and a single hydrogen ship sharing the rest, and Korea's top three yards hold an estimated 70 per cent of global LNG-carrier production.

The same conditions that raise a Korean yard's steel and power bill therefore strengthen the case for the ships it sells, so the net effect turns entirely on product mix: a yard turning out standard bulkers gets the cost pain and none of the relief, and 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, which strands 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, in propulsion. 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 prior year, as demand for dual-fuel and alternative-fuel engines surged, and their operating margins widened accordingly. As owners hedge against future fuel and carbon costs, the premium moves to the propulsion systems only a handful of firms can supply.

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: advanced cryogenic containment for gas carriers and specialist marine engines both sit with a short list of suppliers, so a shock reaches the most profitable segment 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 transition-compliant tonnage 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 the American suspension of Section 301 port fees on China-linked and China-built vessels expires; if it lapses rather than rolls over, a fee schedule starting at 50 dollars a net ton and climbing 30 dollars a year toward 140 by 2028 switches back on, and a ship built in a Chinese yard can become commercially unattractive to an owner trading into the United States regardless of how well it was built.

The other is the International Maritime Organization's Net-Zero Framework vote, deferred once already when 57 states postponed it in October 2025, now due in October 2026, with the earliest entry into force on 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.

The dated stress calendar

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. The near-term risks are diary entries rather than blackouts: two decisions in the autumn of 2026 shape the next order cycle.

Five risks, ranked by how badly they bite

1. Steel cost inside a fixed price. The most immediate and least escapable. Energy-driven plate inflation lands on contracts signed years earlier, and steel is the dominant materials cost. How badly it bites depends entirely on the margin and backlog a yard started with, which, for now, are the best in over a decade.

2. Regulatory timing. Higher than a raw price shock, because it does not merely cost money, it stops decisions. An owner who does not know which fuel will be penalised does not order. Three overlapping instruments, the IMO vote, the EU carbon regime at full bite, and the Section 301 schedule, all resolve between late 2026 and 2028.

3. Demand mix. The order book is a leveraged bet on energy trade. A shock that lifts gas-carrier demand and one that eventually destroys the underlying trade look identical for several quarters, and getting the fuel wrong strands the owner and the specialist yard alike, as the methanol collapse already showed.

4. Supplier concentration. The inputs carrying the most energy risk are also the ones with the fewest suppliers: advanced LNG containment with a handful of licensors, marine engines with a short list of makers, industrial gases with a few regional operators. A single shock arrives through several narrow channels at once.

5. Power price and grid failure. Last, and genuinely lower than it looks. Yards are electricity-intensive but not fragile, and an outright blackout at a major cluster is high-impact but low-probability. The live problem is Korean tariffs grinding upward against contracts that cannot move, and the Chinese tail risk of administrative rationing.

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, which is exactly why the industry can afford to 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 the same market.

It will not break in a single energy shock. A welding bay pauses and restarts; a blast furnace freezes. The 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.