US–China Supply Chain Fragmentation
and Dynamic Capability Reconfiguration

The structural integration of US and Chinese supply chains, which deepened through successive rounds of trade liberalization following China's WTO accession in 2001, is now undergoing a deliberate and accelerating reversal. The initial instrument of this reversal — Section 301 tariffs imposed from 2018 onward — has since been superseded by a comprehensive policy architecture that includes extraterritorial export controls on advanced semiconductors and chipmaking equipment (BIS, 2022; 2023), tightened inbound foreign investment screening through CFIUS, coordinated allied technology restrictions through the Chip 4 framework, and domestic industrial policy instruments of the scale of the CHIPS and Science Act (2022). China's policy response — embedded in its Made in China 2025 program and the dual-circulation strategy — constitutes a structural counterpart: an effort to reduce critical import dependencies while developing indigenous capability in contested technology domains. The symmetry of these policy postures is analytically significant: both states are now explicitly prioritizing technological autonomy over allocative efficiency in GVC design.

From the perspective of international business theory, this trajectory represents a qualitative shift in the nature of the institutional environment governing MNE strategy. The post-1990s international production system was organized primarily around the logic of comparative advantage and transaction cost minimization, with institutional stability treated as a background condition rather than a strategic variable. Geopolitical fragmentation ruptures this assumption. Supply chains across semiconductors, electric vehicles, critical minerals, and advanced materials are being reconfigured not in response to market price signals but pursuant to state-directed objectives of supply security, technological sovereignty, and strategic decoupling. The distinction matters theoretically: when institutional change is the primary driver of environmental discontinuity, the firm-level capabilities required for adaptation are categorically different from those that generate operational superiority within a stable institutional context.

The analytical implications for strategic management theory are specific. First, the efficiency rents generated by highly integrated, just-in-time GVC architectures — rents that accrued precisely from the elimination of redundancy — are being eroded by the introduction of institutionally-mandated slack. Second, the geopolitical realignment of production geography violates the assumption that comparative advantage determines location choice; institutional alignment is now a first-order location variable. Third, and most theoretically consequential, firms whose competitive position rested on operational capabilities — the capacity to execute existing supply chain configurations with maximal efficiency — face structural pressure to develop adaptive capabilities of a fundamentally different character: specifically, the higher-order capacity to sense geopolitical discontinuities, seize alternative resource configurations, and reconfigure organizational architectures under conditions of institutional uncertainty. This distinction, elaborated in the dynamic capabilities framework of Teece, Pisano, and Shuen (1997) and Teece (2007), provides the central theoretical lens for the analysis that follows.

Aggregate trade flow data document a measurable and sustained shift in bilateral trade intensity between the United States and China. The US–China bilateral share of total US goods imports declined from approximately 22% in 2017 to under 14% by 2023 — a contraction of approximately 37% in relative share terms. Concurrent increases in the import shares attributable to Vietnam, Mexico, India, and other ASEAN economies are consistent with two partially distinct mechanisms: trade deflection (the rerouting of Chinese-origin value-added through third-country processing hubs to circumvent tariff and export control exposure) and genuine production relocation driven by compliance requirements, risk diversification imperatives, and incentive-based FDI attraction policies in recipient countries. Disentangling these mechanisms empirically is non-trivial and matters for interpreting the depth of structural decoupling.

In the semiconductor sector, the effects of US export controls — particularly the October 2022 BIS restrictions on advanced logic chips (≤14nm process nodes), high-bandwidth memory, and chipmaking equipment — are visible in customs and industry data. Chinese imports of restricted categories declined sharply post-controls, while domestic Chinese foundries and foreign firms operating in China accelerated capacity expansion in legacy node categories (≥28nm) not subject to restriction. Simultaneously, OECD and UNCTAD FDI data indicate a bifurcation in cross-border investment patterns: greenfield manufacturing FDI into ASEAN, India, and Mexico has accelerated substantially from 2021 onward, while technology-sector cross-border M&A between US and Chinese firms has effectively ceased following CFIUS enforcement intensification. This investment bifurcation reflects, at the aggregate level, the seizing responses of large MNEs navigating the new institutional landscape.

The IMF (2023) provides quantitative estimates of the welfare implications of geoeconomic fragmentation scenarios. Modeling a deep decoupling scenario, the WEO projects long-run output losses of 0.2% to 7% of GDP, with the steepest losses concentrated in small open economies embedded in contested GVCs — precisely those economies whose growth trajectory had depended on GVC integration. These estimates are sensitive to modeling assumptions regarding the speed of adjustment and the availability of alternative trade partners, but they establish an order of magnitude for the systemic costs of the institutional shift underway.

Teece, Pisano, and Shuen's (1997) foundational distinction between operational capabilities and dynamic capabilities provides the theoretical architecture necessary to analyze firm-level responses to supply chain fragmentation. The distinction is precise and consequential: operational capabilities — what Winter (2003) terms "zero-level" capabilities — are the organizational routines that enable a firm to perform a given set of productive activities with maximal efficiency. Dynamic capabilities, by contrast, are the higher-order capacities through which a firm alters its operational capability set: the ability to sense shifts in the competitive environment, seize new resource configurations, and reconfigure organizational structures, knowledge assets, and governance arrangements in response. Critically, operational and dynamic capabilities are not merely different in degree but in kind. An operationally excellent firm — one that has optimized its lean GVC architecture to minimize inventory, reduce lead times, and extract cost efficiencies — has done so by suppressing precisely the organizational slack and routinized experimentation that dynamic capabilities require. This is why supply chain fragmentation does not simply create a problem that existing capabilities can solve with additional effort; it creates a problem that existing capabilities are structurally ill-suited to address.

Teece (2007) elaborates the dynamic capabilities construct into three analytically distinct microfoundational classes: sensing, seizing, and reconfiguring. Each maps onto a specific demand that geopolitical fragmentation places on the firm, and the three classes operate at different temporal horizons and organizational levels, a point that prior applications of the framework to supply chain contexts have underemphasized. Sensing is prospective and environmental-scanning in character; seizing is resource-mobilization in character and operates at the decision-architecture level; reconfiguring is transformational and operates on organizational identity, knowledge structures, and governance systems. A firm may exhibit strength in one class while exhibiting significant deficiencies in another — and it is the weakest link in the sensing–seizing–reconfiguring sequence that constrains overall adaptive performance under conditions of institutional discontinuity.

A critical theoretical corollary concerns the capability trap identified by Levinthal and March (1993): firms deeply invested in the optimization of existing supply chain architectures exhibit a systematic tendency to underinvest in the dynamic capabilities required to transform those architectures. The mechanism is reinforcing. The lean, integrated GVC architecture that generated operational efficiency in the pre-fragmentation institutional environment represents not merely a physical asset configuration but a deeply embedded organizational routine — one whose very success suppresses experimentation, generates high switching costs, and creates powerful internal constituencies for the status quo configuration. Path-dependence of this character implies that the firms best positioned to navigate fragmentation are not necessarily those with the largest established GVC footprint, but those with prior experience managing institutional complexity across discontinuous environmental shifts: firms for whom dynamic capability development has been a prior strategic imperative rather than a novel one.

A second theoretical implication, less developed in the existing dynamic capabilities literature, concerns the qualitative distinctiveness of institutional uncertainty relative to market uncertainty. Market uncertainty — demand volatility, price fluctuations, technological substitution risk — is in principle hedgeable through financial instruments, portfolio diversification, and contractual arrangements. Institutional uncertainty — rooted in the sovereign behavior of states, the exercise of regulatory discretion by administrative agencies, and the evolving boundaries of geopolitically-defined trade blocs — is not contractually mitigable in the same way. Its temporal horizon is longer, its causal structure is more opaque, and its distributional effects across firms are asymmetric in ways that market signals do not reveal in advance. This distinction implies that the sensing component of dynamic capability must, in geopolitically turbulent environments, be extended to incorporate what might be termed political sensing routines: organizational processes for monitoring, interpreting, and integrating geopolitical intelligence into strategic planning cycles. The development of such routines constitutes an adaptive capability investment that existing dynamic capabilities theory has undertheorized and that the current fragmentation context renders urgent.

For multinational firms, the primary strategic imperative is the development of what the dynamic capabilities framework would term geopolitical portfolio sensing and reconfiguring capacity — the ability to design supply chain architectures that maintain operational continuity across multiple fragmentation scenarios without collapsing into the pure redundancy logic that eliminates efficiency advantages entirely. This is analytically distinct from procurement-level dual-sourcing decisions, which represent operational adaptations within a fixed GVC architecture. It requires firms to assess the geopolitical sensitivity of each node in their value chain, identify chokepoints where supply or market access could be subjected to extraterritorial enforcement, and develop organizational routines for rapid supplier qualification in alternative geographies. Critically, the reconfiguring capacity required here involves rebuilding institutional knowledge about unfamiliar regulatory environments — a form of absorptive capacity development that cannot be purchased as a discrete asset but must be accumulated through deliberate organizational investment over time.

For emerging market firms, supply chain fragmentation creates a structural opportunity whose realization is contingent on dynamic capability endowment. The acceleration of manufacturing FDI into Vietnam, India, Indonesia, and Mexico reflects demand for alternative production locations. However, the capacity to convert increased GVC participation into sustained competitive position — through supplier upgrading, absorptive capacity development, and the appropriation of knowledge spillovers from MNE partners — constitutes an adaptive capability challenge of precisely the kind that the fragmentation literature has underspecified. The risk of being incorporated as a low-value-added transshipment node, rather than a genuine beneficiary of production upgrading, is highest for firms whose dynamic capabilities are limited to seizing (accepting incoming FDI) without the reconfiguring capacity needed to embed new knowledge into organizational routines. This distinction between seizing and reconfiguring as analytically separable capability classes has substantive policy implications for emerging market governments designing GVC integration strategies.

For policymakers, the central design challenge is constructing industrial and trade policy instruments that stimulate dynamic capability development at the firm level — particularly the reconfiguring capability class — without generating the static efficiency distortions associated with conventional protectionism. Subsidy programs calibrated solely to cost reduction (direct capital grants for fabrication facility construction, for example) may attract investment without generating the organizational learning, R&D intensity, and institutional knowledge accumulation that constitute the microfoundations of durable dynamic capabilities. Policy frameworks that incentivize inter-firm R&D collaboration, workforce reskilling linked to technology upgrading trajectories, and participation in international regulatory standard-setting are more likely to generate the capability foundations that sustain competitive advantage beyond the initial policy-induced investment horizon. For Korea and ASEAN economies in particular, this implies a theoretical case for FTA institutional designs that preserve supply chain configuration flexibility while building the organizational infrastructure for iterative capability upgrading — a design objective that differs materially from the market access liberalization logic that historically anchored FTA negotiations.