Introduction: When attraction becomes exposure
For several years, Europe’s strategy towards China has been changing in kind. A logic of flows (imports, exports and trade imbalances) has given way to a logic of anchoring. Faced with trade tensions and the desire to protect industrial employment, the European Union and its member states now encourage Chinese companies to establish a local presence through factories and R&D centres. Localisation has become a central instrument of European industrial policy. It is presented as a remedy for vulnerability although, if left unmanaged, it may become one of its vectors.
The shift embodied in the Industrial Accelerator Act is generally presented as a pragmatic response to a fragmented globalisation: produce in Europe rather than import and therefore preserve jobs. The reasoning rests on an implicit assumption: productive anchoring would reduce strategic vulnerability by replacing external dependencies with territorially embedded interdependence. Seen in this light, industrial investment is regarded as a stabilising force, even a source of sovereignty.
However, this European policy raises questions. Localisation is not merely a change in the geography of production; it profoundly alters the nature of industrial and normative ties. By bringing non-European actors into the heart of Europe, inside our established ecosystems, it creates a tighter bond, one that is hard to reverse and carries long-term dependencies. Industrial anchoring therefore makes dependence less visible but, over time, more constraining. This dependence is silent because it is marked neither by crisis nor by any identifiable rupture. It gradually takes root in the technical and contractual layers of the productive system that attract little attention, until it is accepted as the very condition of cooperation.
This article builds on the work of Henry Farrell and Abraham Newman on ‘weaponized interdependence’, which shows how economic and technological interdependencies can become instruments of structural power.¹ It offers an industrial and territorial reading of that idea, applied to Sino-European partnerships and productive localisation strategies. It is precisely this grey area that the article seeks to explore, in order to shed light on the implicit doctrine emerging behind the choice of productive localisation as a response to industrial vulnerabilities.
The aim is to bring into view the rarely stated principles that already shape China’s anchoring in Europe: what we accept sharing, what we believe we can make conditional, and what we assume we can ring-fence. The challenge is to understand how Sino-European partnerships are redrawing our industry and our room for decision-making.
Through the example of the automotive industry, followed by the counterpoint of aerospace, the article seeks to estimate the threshold at which the industrial anchoring of non-European actors ceases to be a lever of sovereignty and becomes a source of strategic vulnerability. Above all, it asks whether Europe now has the conceptual and institutional tools to govern this transformation rather than endure it.
This question is especially salient at a time when the United States has redefined the conditions under which foreign industrial actors may anchor themselves on its soil, turning conditionality into an explicit instrument of industrial policy. Europe, for its part, has begun to frame the issue through the Industrial Accelerator Act, but this act was conceived primarily as a tool for accelerating decarbonisation and simplifying administration rather than as a framework for governing non-European dependencies. Conditionality remains marginal, and the anchoring of non-European actors in subsidised value chains is not addressed head-on.
I. Partnership as an instrument of power
A. When cooperation ceases to be neutral
For a long time, industrial partnerships in Europe were understood as a tool for addressing clearly defined economic constraints. They made it possible to share costs or pool the risks associated with innovation. In this model, cooperation remained fundamentally reversible. It was a calculation of efficiency, not a long-term strategic commitment.
In today’s world, that analytical framework is no longer valid.² In industries built around complex systems, cooperation goes further than simply adding together independent capabilities by interweaving architectures and technological trajectories. To cooperate is now to co-produce the rules of the game within which future innovation will unfold. As a result, partnership ceases to be a neutral tool and becomes a space in which positions of power are allocated, often quietly.
This transformation remains widely underestimated in Europe. Cooperation is still assessed using visible criteria such as the scale of investment or the number of jobs created. Yet the essentials lie elsewhere, in the definition of technical interfaces and standards. These intermediate layers, unglamorous but decisive, shape future dependence when they slip beyond European control.³
B. China’s strategy as the patient accumulation of leverage
The distinctive feature of China’s strategy lies less in the scale of the resources deployed than in their systemic coherence. Industrial partnerships are conceived as successive stages in a long-term trajectory: learning from Western technologies, gradually integrating into global value chains and, eventually, acquiring the ability to shape standards.
Taken in isolation, each partnership appears balanced, sometimes even favourable to Western interests. It is their aggregation over time that produces a tipping effect. The semiconductor industry offers the clearest demonstration because the outcome is now visible. For two decades, China methodically forged partnerships, joint ventures and technology-transfer agreements with leading Western players: IBM licensed the Power architecture to a Chinese consortium; AMD established THATIC with Tianjin Haiguang around x86 processors; ARM ceded operational control of ARM China; Qualcomm partnered with Guizhou province on server chips; and Intel invested in Tsinghua Unigroup. Separately, each of these agreements seemed balanced, and sometimes even advantageous to the Western partner, which saw in it a route into the Chinese market. Their patient accumulation has produced the trajectory now before us: SMIC is able to etch at 7 nanometres, Huawei designs its own processors through HiSilicon, and YMTC competes with Samsung and SK Hynix in 3D NAND memory. None of these partnerships was decisive on its own. It was their articulation within a long-term strategy, supported by the Made in China 2025 plan, that enabled the emergence of a largely autonomous ecosystem. The example is all the more instructive because semiconductors, unlike cars, already show the end point of the dynamic: what Europe’s automotive trajectory suggests, the semiconductor trajectory has already delivered.⁴
Huawei’s announcement of 25 May 2026, made on the sidelines of the International Symposium on Circuits and Systems in Shanghai, illustrates the culmination of this trajectory. He Tingbo, head of the group’s semiconductor division, said the company would be able by 2031 to produce chips etched at 1.4 nanometres – the node TSMC is targeting for 2028 – without using ASML’s extreme-ultraviolet lithography machines, access to which is barred by US sanctions. The announced break is not about etching finesse itself, but about the design paradigm: “Instead of optimising space,” she said, “designers optimise the time taken for communication between the different elements of a chip.” Beyond the timetable, which remains to be confirmed, the statement marks a shift. Instead of
seeking cooperation primarily to gain access to Western technological building blocks, China is seeking the capacity to propose its own architectures and, over time, impose its own standards. Dependence is being reversed⁵, with Chinese actors using cooperation as a mechanism for accelerated learning. Given that China has already reached a level of excellence in most sectors, its aim has largely shifted from acquiring the technologies themselves to absorbing methods, standards and skills that are hard to formalise and protect, and then reinvesting them in proprietary architectures progressively freed from their Western source.
This dynamic is reinforced by a deeply asymmetric relationship with time. Whereas European companies still tend to think in terms of a programme or an investment cycle, their Chinese competitors embed partnerships within much longer horizons. This continuity is supported by a tight alignment between industrial strategy, financing and public policy.⁶ For Chinese manufacturers, partnership serves as a way of shaping the playing field and gradually shifting technological or decision-making centres of gravity, without direct confrontation or visible rupture.
C. Europe’s blind spot: partnerships without a doctrine
Faced with this coherent Chinese strategy, Europe has so far proceeded without any real doctrine for governing industrial partnerships. Decisions are taken case by case, at national level or by companies themselves, with no consolidated assessment of trajectory effects.⁷ This fragmentation is part of an intellectual legacy from a world in which open markets and economic integration were seen as intrinsically stabilising. Without quite realising it, Europe continues to treat a power problem as a question of attractiveness.
In a world where interdependence has become an instrument of power, Europe’s approach shows its limits. Partnerships can create functional dependencies long before they are politically identified as such. Territorial anchoring makes these forms of cooperation economically and socially legitimate and therefore hard to reverse. Dependence is produced from within, in the name of competitiveness and attractiveness. Localisation, initially conceived as a bulwark, becomes the very mechanism through which dependence is made socially acceptable, economically rational and politically costly to challenge.
This is where the central strategic risk lies. So long as Europe continues to treat Sino-European partnerships as isolated economic transactions, it will expose itself to a gradual loss of capacity to arbitrate its own industrial trajectories. Recognising that international economic relations are now governed above all by logics of power is the essential precondition for any serious reflection on industrial localisation and European sovereignty.
Europe, engaged in a brutally fast technological transition, has incorporated new critical building blocks without an equivalent governance framework. It is here that productive localisation, presented as a solution, most clearly reveals its potential to generate dependence.
II. The European car industry as a laboratory for silent dependence
A. Industrial localisation as an apparent solution
Europe’s car industry now sits at the heart of the localisation strategy pursued by the European Union and its member states.⁸ Faced with the rise of Chinese manufacturers, the dominant response has not been exclusion but anchoring. Replacing import flows with local production, attracting factories in the hope of attaching R&D centres to them, preserving industrial jobs and integrating these players into the European ecosystem have become explicit goals, both for governments and for some carmakers seeking to transfer certain plants to Chinese manufacturers.
The agreement signed by Stellantis and Dongfeng in May 2026 offers a telling example. While the Chinese component provides for the production of Peugeot and Jeep models in Wuhan, the European component entrusts Stellantis’s Rennes-La Janais plant with assembly and distribution on the continent of a Voyah model, Dongfeng’s premium brand. The French site, which now assembles only the Citroën C5 Aircross, thus sees part of its industrial future tied to a Chinese manufacturer. Localisation preserves jobs and productive capacity, but it also embeds in the European ecosystem a vehicle whose technological architecture is still designed and governed outside Europe. For Dongfeng, the move also makes it possible to bypass customs duties on electric vehicles imported from China. The case captures the broader mechanism: localisation solves a short-term problem (underused capacity) at the price of an architectural dependence that is difficult to reverse.
Not all Sino-European cooperation falls into the same category, however. Horse Powertrain, the joint venture structured on a 50-50 basis between Renault and Geely in May 2024 and subsequently opened up to a third shareholder, concerns internal-combustion and hybrid powertrains, a segment in the downward phase of Europe’s investment cycle, yet still important for emerging markets and the global hybrid segment. That explains the converging interest of the two partners. This type of partnership, built around assets whose strategic value to Europe is declining and supported by balanced governance, illustrates the possibility of controlled cooperation: Europe uses mature industrial know-how to absorb part of the fixed costs of a technological transition it is otherwise pursuing through electric vehicles. The strategic question is different when partnerships concern emerging architectures (software platforms, battery-management systems and data environments) that will shape the next generation of vehicles. It is this shift in the industrial centre of gravity that must now be examined.⁹
Batteries already provide a more advanced demonstration. Envision AESC’s plant in Douai, opened in 2024 next to Renault’s ElectriCity hub, now supplies the electric Renault 5 and 4 and is, according to the French manufacturer itself, the first genuinely operational gigafactory on French soil. The partnership is rightly presented as an industrial success: it has localised critical productive capacity and created several thousand jobs in Hauts-de-France. Yet the cells’ technological architecture, chemistry choices, thermal-management standards and software control environments remain largely defined within the Envision group, whose parent company is Chinese and whose main R&D centres are outside Europe. CATL’s plant in Arnstadt, near Erfurt, which has supplied BMW, Stellantis and several other European carmakers with lithium-ion cells since 2023, follows the same logic: the industrial presence is French or German, but the cell architecture and technology road maps are defined in Ningde.¹⁰Here, architectural dependence is already in place; it remains to acknowledge it.
The localisation strategy rests on an intuitive logic: producing on European soil reduces the vulnerability associated with large-scale imports; job creation is seen as proof of reciprocity; and a local foothold is expected to align the economic interests of Chinese investors with those of host regions. At first sight, localisation therefore looks like an answer to fragmented globalisation.
But if one accepts that cars have become a systems industry, one must also accept that strategic value has shifted from the factory itself to the technological architectures that govern it. Localising production does not necessarily mean controlling those architectures. Indeed, physical presence can conceal a far deeper functional dependence, embedded in choices of platforms and standards.
B. Where the centre of gravity is really shifting
In today’s car industry, power lies as much in the ability to define the conditions under which vehicles evolve as in the ability to produce them. Electronic architectures, battery-management systems and assisted-driving environments, to name only a few, now structure the entire value chain. In electric vehicles, the battery-management system and data environments determine product evolution far more than final assembly does.
This is precisely where real dependence is at stake. Industrial partnerships and local plants can preserve jobs and productive capacity while, de facto, transferring architectural power. When an ecosystem is organised around platforms designed elsewhere and performance standards escape local control, Europe’s ability to steer its own industrial trajectories is diminished.
This shift is all the harder to perceive because it is accompanied by no visible rupture. The factories operate and investment continues. Public decision-makers find material for positive announcements. Meanwhile, dependence is being installed in the intermediate layers of the productive system. One day, replacing a technology or withdrawing from a partnership will be difficult without incurring high costs. This dependence is produced gradually, through integration, rather than an abrupt shock.¹¹
C. European fragmentation and the acceleration of dependence
Europe’s fragmentation amplifies this dynamic. Strategies for welcoming Chinese investment differ markedly from one member state to another: some prioritise immediate attractiveness and rapid job creation, while others are more cautious. But there is no common framework strong enough to alter the overall trajectory. Seen from Beijing, Europe can appear as a succession of leaders taking turns to extol their country’s appeal beneath the impassive gaze of the Chinese Sphinx.
This heterogeneity creates the conditions for Chinese actors to integrate unevenly across Europe’s industrial space. Plants are located where regulatory, fiscal or social conditions are most favourable, without any consolidated assessment of their systemic effects at Union level.¹² Localisation then becomes a fait accompli, difficult to reverse not only for the host state but for the internal market as a whole. The contrast with the United States exposes Europe’s blind spot. The Inflation Reduction Act provisions on Foreign Entities of Concern in effect exclude Chinese actors from subsidised value chains in batteries and electric vehicles, whereas Europe integrates them without equivalent conditionality.¹³
In this context, the car industry appears as an advanced laboratory for a broader dynamic. Rapid technological transition, the absence of a normative authority comparable to those found in other systems industries, and competition between territories have allowed structural dependencies to form before they were even fully identified. The localisation of factories, presented as a solution, thus reveals its limits. It may strengthen resilience in the short term while weakening strategic room for manoeuvre over the longer term.
This contrast invites us to look at another great European industry, one that is also globalised and highly capital-intensive, and that has long faced comparable interdependencies but without losing its capacity for control: aerospace.
III. Aerospace as a counterpoint in governance
A. A globalised industry that has not lost control
European aerospace is one of the most globalised industries in existence. Its value chains span continents, its suppliers are spread across several jurisdictions and its markets are deeply international. Yet so far, it has not suffered a comparable loss of strategic room for manoeuvre.¹⁴ This distinctiveness deserves attention as it is neither natural nor spontaneous: it is the result of patient institution-building, often born of crises and explicit political choices.
The key distinction lies in the organisation of the industry’s interdependence, not in any lower exposure to international markets or defensive closure. Indeed, aerospace has long been understood as a critical system rather than a simple assembly industry: safety, reliability and operational continuity determine access to the market. This logic led early on to explicit governance of architectures, standards and partnerships.
Within this framework, instead of erasing Europe’s power to set the terms of integration, globalisation made that power more demanding. Openness was accompanied by institutions able to define the rules for integration, certification and the evolution of the industrial system. International cooperation is intense, but it operates within a stable normative framework that limits the formation of irreversible dependencies.
B. Standards, certification and quiet power
At the heart of this governance lies the central role of standards and certification. In aerospace, access to the market depends on compliance with safety and traceability requirements set by recognised authorities, in addition to economic or technological competitiveness. This normative power, often perceived as merely technical, is in fact a major strategic lever.
Certification serves to validate a finished product, but it also shapes upstream architectures and value chains. It imposes integration standards that make some dependencies acceptable and others unworkable. In this framework, industrial partnerships are bound by rules designed to ensure control over critical interfaces and the ability to withdraw in the event of a strategic rupture.
The centrality of the normative framework explains why aerospace has been able to absorb complex forms of international cooperation without losing its overall coherence. Power lies in the capacity to define the conditions under which the industrial system may be accessed and evolve, rather than in the isolated ownership of assets. Ultimately, dependence is governed, not endured.¹⁵
C. What the car industry failed to institutionalise
The contrast with the car industry is striking. Caught up in an exceptionally rapid technological transition, it has incorporated new critical building blocks without an equivalent governance framework. Electrical and software architectures imposed themselves faster than the institutions capable of setting their rules.¹⁶ Partnerships multiplied without the conditions for reversibility or normative control being explicitly set.
Where aerospace built institutional guardrails before dependencies became irreversible, the automotive industry has often done the opposite. Cooperation has been conceived as a short-term response to technological and competitive urgency, with trajectory effects pushed into the background. This lag explains why the establishment of factories can now produce dependencies in the automotive industry, while aerospace maintains a capacity for arbitration. The difference is a corollary of the order in which such openness and safeguards were institutionalised: in aerospace, the rules governing interfaces preceded the strengthening of interdependencies, not the other way round.
The point is not to transplant a sectoral model mechanically from one industry to another. That would be illusory, given how profoundly their constraints and markets differ. But the lesson is clear: industrial sovereignty depends less on the degree of openness than on the capacity to establish governance rules before trajectories become fixed. Where such governance is lacking, dependence is built quietly.
This counterpoint illuminates the core of Europe’s problem. Is Europe capable of governing the industrial anchoring it encourages? That is the question a broader reflection on the principles of a European doctrine of localisation and dependence must answer.
IV. Governing without closing: towards an implicit European doctrine
A. What Europe is trying to achieve through localisation
Europe’s industrial localisation strategy pursues legitimate goals. It seeks to preserve employment, secure value chains, reduce exposure to external shocks and support the ecological transition. In a world where trade flows are increasingly politicised, productive anchoring appears to be an instrument of economic and social stabilisation.
Yet this strategy rests on an assumption that is rarely made explicit. It assumes that localising industrial capacity is enough to reduce strategic vulnerability by aligning the interests of foreign investors with those of host territories. The factory, the R&D centre and the local ecosystem are thus invested with an implicit political function: they are expected to turn external dependence into managed interdependence.
That transformation is neither automatic nor guaranteed. Localising an activity does not necessarily mean controlling the architectures that structure it, or steering the technological trajectories that govern it. When fundamental choices about platforms and standards are made elsewhere, territorial anchoring can strengthen short-term resilience while weakening long-term decision-making capacity. Localisation then produces a paradoxical effect: it makes dependence socially acceptable and politically costly to challenge.
B. What Europe does not yet govern
This paradox reveals a deeper gap: Europe does not yet have an explicit doctrine for governing industrial dependencies. The instruments exist, but they remain fragmented. Attractiveness policies, competition frameworks, investment-screening mechanisms, standard-setting strategies and industrial initiatives move forward in parallel, without being systematically linked to a view of long-term trajectories.
The absence of a unified framework leads attention to settle on visible variables. Investment volumes, the number of jobs created and the geographical location of sites become central indicators, while the decisive levers lie elsewhere. Control over critical interfaces, software evolution, the capacity to influence standards and the securing of data remain largely outside the scope of routine political assessment.
In this context, dependence appears as a gradual erosion of the capacity to arbitrate. Each partnership, taken in isolation, seems controlled, yet their accumulation creates a system in which certain decisions become impracticable because of the economic and social costs they would impose on a territory or state. Dependence is thus produced by a failure of governance.¹⁷
C. Towards sovereignty through control of trajectories
This observation invites a reframing of European industrial sovereignty. Restricting it to the reshoring of capacity or the closing of markets would be reductive. Instead, it must rest on the ability to choose acceptable dependencies, limit their irreversible effects and preserve room for manoeuvre over time.
From this perspective, sovereignty is exercised not only through ownership of assets, but through control over trajectories.¹⁸ In other words, industrial sovereignty is predicated on the art of choosing, ordering and governing dependencies in a way that preserves credible long-term room for decision-making. It requires a distinction between what must be ring-fenced, what can be shared and what must remain conditional. It also requires particular attention to the architectures, standards and governance frameworks that structure industrial ecosystems, far more than to visible physical sites alone.
Europe’s challenge, therefore, is to reconsider industrial localisation, placing it within a more demanding conceptual framework. Without such a framework, productive anchoring risks turning a strategy of resilience into a mechanism of dependence. With it, localisation can become a lever of open sovereignty, able to combine international cooperation with autonomous decision-making capacity.
When localisation ceases to be a simple economic tool and becomes a structuring political fact, it calls for a change of perspective. The central question thus shifts from whether industrial investment should be attracted to whether Europe can govern its long-term effects. Only on that condition can the establishment of foreign factories strengthen European sovereignty.
Conclusion
Europe’s industrial localisation strategy was born of a desire to regain control over value chains weakened by fragmented globalisation. Attracting investment to rebuild productive capacity and anchor employment on European soil have become central objectives, presented as responses to a more uncertain international environment. From this perspective, localisation is often conceived as a form of protection. It is supposed to turn external dependence into stabilised interdependence.
The analysis set out here calls for greater caution. Localisation is not, in itself, a remedy for dependence. Indeed, in systems industries, it may even become a lasting vector of dependence when technological architectures and standards escape the control of the Union and its member states. Where once such dependence revealed itself through the visible loss of assets or abrupt breaks, now it is built through gradual integration, through the accumulation of technical and organisational choices that make reversibility ever more costly.
However, Europe’s entanglement with Chinese industrial investment is already too deep to now reject it outright, – doing so would not be in Europe’s interest – and a solution involving the closure of markets would be illusory. Instead, a more demanding imperative emerges: the foundation of industrial sovereignty has shifted from the material ownership of factories to the capacity to steer the systems that govern them. Where that capacity is lacking, industrial establishment can weaken what it claims to strengthen.
Europe’s task, then, is to equip itself with the conceptual and institutional tools needed to govern the dependencies it helps to produce. The aim is less to reduce interdependencies than to master their effects; less to limit cooperation than to define the conditions under which it can be reversed.
When localisation becomes dependence, it is because the trajectory created by foreign investment is insufficiently governed. Conversely, investment embedded in a controlled normative and architectural framework can become a lever of open sovereignty, compatible with international cooperation and autonomous decision-making capacity.
Without such clarity, Europe risks discovering too late that the anchoring it encouraged has, without explicit debate, redrawn the boundaries of its own autonomy. This is where the future of European industrial sovereignty will now be decided: not in rejecting anchoring, but in understanding how it binds Europe. In practical terms, such sovereignty rests on three requirements. Ring-fencing the functions and architectures that cannot be exposed without an irreversible loss of capacity to arbitrate. Sharing, within explicit frameworks, what can be shared without surrendering the power to define technological and normative trajectories. Making access conditional to the European market on control over the critical interfaces and standards that structure industrial ecosystems.
These three requirements can draw on instruments the Union already possesses, provided it is prepared to deploy them more systematically. The EU Framework for Screening Foreign Direct Investment¹⁹, in its revised form, provides a basis for identifying what must be ring-fenced, provided its scope is widened to include indirect stakes and forms of architectural control. Important Projects of Common European Interest (IPCEIs)²¹, like the Foreign Subsidies Regulation (FSR)²², which entered into force in 2023, offer a lever for making access to public funding and markets conditional on effective control over critical interfaces, in much the same way as the United States uses Foreign Entities of Concern under the Inflation Reduction Act. European standardisation, finally, remains the most underused field, even though it is the most decisive lever for sharing without surrendering control: the experience gained in civil aerospace certification shows that sustained European involvement in standard-setting bodies is compatible with dense international cooperation. These instruments exist. What remains to be built is their articulation around a common doctrine.
Localisation becomes a lever of open sovereignty only when it is conceived around these red lines, rather than as a mere instrument of attractiveness or flow management.
Frédéric Recordon – Partner, Accuracy
When Localisation Becomes Dependence – European industrial sovereignty and China’s localisation strategy