Charted: China Builds 55% of Merchant Ships — but Only 6 Shops Forge Nuclear-Class Heavy Metal

Software scales on racks. Physical capital stock still scales on yards, dry docks, final-assembly lines, and ultra-heavy forges. The core industrial question for 2026 is not “who announced a factory,” but who can still cut, weld, float, and forge the things that take decades to replace: merchant hulls, large jets, reactor-class pressure vessels, and the dock capacity that keeps those fleets alive.

This research cross-walks five builder bases that rarely share a slide: merchant shipbuilding GT deliveries, VLCC-capable dry docks, large-jet final-assembly sites, ultra-heavy nuclear-class forges, and crude steel. Companion deep-dives already map the ship and aircraft slices in detail — global shipbuilding GT delivery concentration and commercial aircraft final-assembly geography. Here the point is the cross-sector pattern: East Asia dominates floatable steel; North America still dominates large-jet handovers; ultra-heavy forging is a six-shop oligopoly; and owning demand is not the same as owning the builder base.

The cross-sector scoreboard

Toggle the dashboard’s Sector shares panel across those five lenses. The leader changes. That is the whole thesis: “industrial capacity” is not one HHI. A country can own steel and shipyards and still lack an RPV-class forge — or host half of jet handovers while building almost no merchant GT.

Shipyards: the Asia trio is the world

UNCTAD’s Review of Maritime Transport 2025 puts China at 54.6% of 2024 merchant GT deliveries, Korea at 28.0%, Japan at 12.6%. The Asia trio clears 95.2%. Europe’s aggregate merchant print is under 1%; the United States is a rounding error at ~0.04%. That is not a cyclical soft patch. It is a multi-decade migration of the yard map, shown in the dashboard’s Ship share path from Japan’s 1980s peak-era dominance through Korea’s ascent to China’s >50% milestone in 2023–24.

Dry docks tell the same story in concrete. On a tracked inventory of VLCC-capable docks, China holds roughly three-fifths, Korea nearly a fifth, Japan high single digits. Repair and naval-specialist yards in Europe and the US still matter for cruise ships, carriers, and submarines — Newport News, Fincantieri, Meyer — but they are not the merchant float machine. Sort the Yard / dock map by large dry docks and the merchant complexes in China and Korea sit at the top of the relative capacity index.

Aircraft FALs: a different continent wins

Open Aircraft FALs. The same physical-capital question applied to large jets flips the map. Attributed 2025 handovers concentrate on a handful of campuses: Renton 31.7%, Hamburg 19.0%, Toulouse 15.2%. The top three sites hold 66%; US final-assembly campuses still clear 54.5% of the large-jet pool even though Airbus spreads A320 family volume across Europe, China, and Alabama.

That is why industrial-policy talk that treats “shipyards and jet factories” as interchangeable East Asian stories is wrong. Merchant GT is an East Asian production network. Large-jet handovers remain a US + Northwest Europe hangar story, with Tianjin and Shanghai as growing but still minority nodes. Our geography companion post walks the site ladder in more depth; the research takeaway here is simpler: builder power in aviation is site-concentrated and Atlantic-heavy, not a mirror of UNCTAD’s ship table.

Build vs own: Greece floats fleets China welds

Switch to Build vs own. Beneficial ownership of the merchant fleet and the geography of deliveries diverge sharply. Greece owns on the order of 16% of world tonnage while building almost none of it. China builds ~55% of new GT while owning closer to 14%. Korea builds far more than it owns. The United States owns a modest fleet share and builds essentially none of the commercial merchant pipeline.

Aviation and steel land closer to the 1:1 line — US FALs and US airline demand are not as mismatched as US shipyards and US cargo ownership — but the shipping scatter is the cleanest illustration of the theme’s meta-question. Capital stock can be owned in Piraeus and fabricated in Jiangsu. Policy that only tracks fleet flags or airline registries will misread where the scarce welding capacity sits.

Ultra-heavy forges: six shops, zero US RPV-class members

The Heavy forges panel is the scarcest layer. Reactor-pressure-vessel-class and comparable ultra-heavy components still depend on a tiny set of shops: Japan Steel Works (Muroran), Doosan Enerbility (Changwon), China First Heavy Industries, Shanghai Electric’s heavy forge complex, Framatome / Le Creusot, and Sheffield Forgemasters as a heavy-vessel peer. Equal-weight across that tracked six-shop set, Japan and China each take about a third of shop count; Korea and Europe split the rest. North America has no shop in this RPV-class inventory.

That matters for nuclear new-build timelines, naval propulsion, and any industrial strategy that assumes “steel capacity” implies “forging capacity.” Crude steel at 53.8% Chinese production share is abundant relative to ultra-heavy forging. The bottleneck is ingot size, heat-treatment bays, and certified nuclear-quality process control, not blast furnaces.

What the sector toggles reveal about risk

Put the five sector bars side by side in your head:

1. Ship / dock shock hits East Asia firsta Chinese yard slowdown or Korean LNG-carrier slot squeeze moves global floatable capital stock immediately.
2. Jet shock hits Renton, Hamburg, and Toulouseone campus can be a third of world large-jet handovers.
3. Forge shock hits a six-address phone booklead times measured in years, not quarters.
4. Steel shock is broader but still China-weighted; the rest-of-Asia and Europe/NA tails are thicker than in shipbuilding.
5. Own-vs-build mismatch means financial claims on fleets and aircraft can be geographically far from the welders and riveters that replace them.

Fiscal and industrial-policy packages that subsidize “manufacturing” without distinguishing these layers will buy the wrong bottleneck. Pair this map with fiscal-industrial policy when the question is which incentives actually touch yards and FALs rather than generic plant investment.

Who is exposed — and who still has optionality

Exposed: shippers and energy traders who assume dry-dock and newbuild slots are fungible outside China/Korea; airlines and lessors whose delivery slots cluster on one FAL campus; nuclear and naval programs that treat RPV-class forging as a competitive market; US merchant-marine ambitions that confront a 0.04% commercial GT delivery share.

Relative optionality: Airbus’s multi-continent FAL network as a hedge inside aviation; Korea’s high-value LNG and complex hull niches inside a China-led GT world; Europe’s cruise/naval specialists and Le Creusot forge as narrow but real capability islands; Japan’s remaining yard and JSW forge footprint as a quality-and-certification hedge even as GT share fades.

What would rewrite the map: a sustained US or European merchant newbuild renaissance large enough to move GT shares out of the noise; a COMAC ramp that lifts Shanghai well above ~1% of large-jet handovers; a new Western RPV-class forge or certified expansion that breaks the six-shop club; or a Chinese dock/yard consolidation shock that forces orderbooks toward Korea and Japan at scale.

Caveats and methodology

• Sectors are not additive. Do not average ship GT share with FAL share into a fake “overall industrial capacity” index.
• Dry-dock and forge counts are tracked inventories, not a complete global registry. Coverage favors large, disclosed complexes.
• Yard capacity index is relative within the dashboard’s samplenot absolute GT or compensated CGT.
• Aircraft site shares for multi-FAL A320 volume are line-estimated in the companion geography post; Renton’s lead is more robust than ranks 2–5.
• Aviation “own” shares on the scatter are demand/fleet proxies, not UNCTAD-style beneficial-ownership statistics.
• Ultra-heavy forge list is capability-class, not revenue-weighted; equal-weight shop shares overstate tiny peers and understate JSW/Doosan throughput.
• Steel production ≠ heavy fabrication. Plate, sections, and forgings are different industrial systems.
• Naval and cruise yards are strategically important but deliberately separated from the merchant GT leaderboard.

The shareable takeaway

Physical capital stock is still built where the yards, docks, hangars, and forges are — and those maps disagree. China delivers 54.6% of merchant GT and anchors ~62% of tracked VLCC-capable docks; the Asia ship trio clears 95%. Renton alone finishes 31.7% of large-jet handovers; US FALs still hold a majority. Only six shops sit in the tracked nuclear-class forge set, none in the United States. Owning fleets or airlines does not own the welder. Watch sector-specific builder shares, not a single “manufacturing” headline, to see who can still expand the physical capital stock.