IN Brief:
- Replicating China-linked supply chains could require $13.7tn in the US, $9.1tn in the Eurozone, and $800bn in the UK.
- Manufacturing, mining, and power and utilities account for almost $13tn of the estimated requirement.
- Higher capital expenditure and prices make selective localisation more likely than full decoupling.
EY-Parthenon has estimated that replicating China-linked supply chains across the United States, Eurozone, and United Kingdom could require approximately US$23.6 trillion of investment by 2050.
The potential requirement comprises US$13.7 trillion in the US, US$9.1 trillion in the Eurozone, and US$800 billion in the UK. Replacement factories would account for only part of the total, which also includes research and development, software, advanced manufacturing, transport networks, supplier ecosystems, infrastructure, and workforce skills.
Manufacturing, mining, and power and utilities represent almost US$13 trillion of the estimated investment because of their exposure to Chinese components, processed materials, production capacity, and upstream supplier networks. Machinery, electronics, and automotive businesses could face particularly large increases in capital expenditure.
Decoupling would develop gradually rather than through a single investment cycle. Early measures could concentrate on critical products, exposed sectors, and strategic nodes, but the required expenditure would rise materially as localisation moved into more complex and deeply integrated value chains.
Government funding on the scale modelled by EY-Parthenon could increase public deficits across the US, UK, and Eurozone by close to one percentage point of gross domestic product annually through 2050. If businesses absorbed the investment instead, capital expenditure in several industrial sectors could reach twice its present level.
Mats Persson, EY-Parthenon UK macro and geostrategy leader, said geopolitical and economic forces were pulling in opposing directions, producing a more fragmented form of globalisation in which economic nationalism would coexist with continued international trade.
“Western economies are already managing capital-intensive transitions across energy, technology, defence and infrastructure; adding full East-West supply chain decoupling without clarity on who bears the cost risks proving unaffordable,” he said.
Industrial ecosystems resist simple relocation
The estimate distinguishes between replacing a particular supplier and recreating the industrial environment surrounding it. A component can sometimes be dual-sourced within a manageable period, while the materials, specialist machinery, tooling, technical knowledge, testing capacity, logistics links, and lower-tier suppliers supporting its production may take years to reproduce.
China’s cost advantage does not rest solely on labour. Dense clusters of processors, component manufacturers, ports, toolmakers, engineers, and equipment suppliers allow businesses to source, modify, test, and scale products within relatively compact networks. A new Western factory may remain dependent on that ecosystem even after final assembly has moved.
Factory prices for some Chinese components are estimated to be between 20% and 100% lower than available Western alternatives. Partial relocation could therefore lift long-term price levels by one to two percentage points, with the cost distributed between company margins, taxpayers, and final customers.
Selective de-risking offers a more credible route than uniform reshoring. Chinese supply chains are themselves widening supplier bases and increasing digital control, creating networks that remain globally connected while reducing dependence on individual sites, lanes, or counterparties.
The appropriate response will vary by product and consequence. A low-cost component may justify substantial resilience investment when its absence can stop a high-value production line, while a more expensive item may remain suitable for global sourcing when alternative suppliers and transport routes are readily available.
Procurement teams therefore need visibility beyond their immediate contractual supplier. Concentration may sit within a sub-tier processor, proprietary tool, raw material, testing facility, or port rather than the company issuing the invoice.
Critical minerals demonstrate the scale of that challenge. New mines can expand extraction, but refining, chemical processing, component manufacturing, permitting, environmental controls, and technical skills must develop alongside them. Fresh capacity must then compete commercially with incumbent producers operating at established scale.
Localisation can also create additional logistics demand rather than simply shortening the chain. New plants require inbound materials, utilities, maintenance inventory, warehouses, workforce transport, and outbound distribution, while domestic roads, railways, ports, and power networks may need investment to support the relocated activity.
A mixed sourcing model is consequently more likely to emerge, combining domestic production for critical items, regional sourcing for high-volume goods, dual supply for constrained components, strategic stock for long lead-time materials, and global procurement where the cost advantage remains substantial and the risk manageable.
Scenario planning becomes more important within that model because tariffs, export controls, shipping disruption, energy availability, and political intervention can alter the economics of a route or supplier quickly. Capital projects expected to operate for decades cannot be justified against a single assumed trade environment.
The US$23.6 trillion figure places a hard economic boundary around political calls for comprehensive decoupling. Companies can still reduce their most serious dependencies, but duplicating an entire China-linked industrial system would compete directly with investment already required for energy transition, defence, digital infrastructure, and ageing public assets.
A more fragmented supply network is therefore likely to retain substantial global trade while adding regional capacity, alternative suppliers, inventory, and transport options around selected risks. The result will be more expensive and operationally complex than the leanest pre-disruption model, but considerably less costly than attempting to reproduce every existing value chain at home.
