Big Tech Digital Sovereignty Conflicts and the Geopolitics of Data Infrastructure: European and North American Cases and Middle Power Responses
Executive Summary
Dependence on Big Tech infrastructure is transforming into a trade and security risk, as local opposition movements against data centers in France, Austria, and Finland converge with Canadian and European responses to the U.S. CLOUD Act. As ECB President Lagarde has pointed out, Europe’s dependence on cloud services could potentially function as leverage for the United States in negotiations over tariffs and digital taxes. However, the baseline trajectory for the next one to two years is expected to be a pattern of parallel but unaligned developments, where sporadic resistance from local politics and data sovereignty discourse from national governments fail to converge into substantive policy implementation (probability: 55%). For South Korean government and corporate actors, European attempts to transition to domestic cloud services could simultaneously create new export opportunities and entry barriers. This necessitates quantifying local political risks in potential investment destinations during the due diligence phase and pursuing a parallel site selection strategy that leverages regulatory variations among U.S. states. The Ministry of Trade, Industry and Energy and the Ministry of Foreign Affairs should prioritize establishing a low-cost, continuous monitoring system, while also exploring channel diversification to prepare for the possibility that data sovereignty discourse could be repurposed as a weapon in trade negotiations.
I. Situational Analysis
The Threat to “Digital Sovereignty” from Big Tech and the Proliferation of National Responses
1. Background and Developments
The flashpoint for the opposition movements against data centers was local politics in small European towns. In the village of Étrechet near Châteauroux in central France, independent mayor Florence Laurent resigned on September 15 by returning her tricolor sash to protest Google’s plans to build a massive data center [1]. This case, where local backlash against the construction of a mega-facility in the middle of farmland suffering from declining wheat yields led to a mayor’s resignation, is a notable example [1]. In Kronstorf, Austria, an opposition movement against Google’s data center construction plans has similarly spread beyond the local community [13]. In Finland, the data center debate entered a new phase after Google announced a €13.1 billion investment in three locations: Kajaani, Valkeala, and Muhos [8]. According to a report by Yle, anti-data center sentiment is growing simultaneously on both sides of the Atlantic, with opposition now belatedly forming in Finland as well [8].
This local backlash is not limited to a simple NIMBY phenomenon. In a commentary, the Swiss newspaper NZZ characterized Silicon Valley’s “doomsday scenarios” as a product of political calculation [5]. NZZ argued that “politics and the economy would be better off countering Big Tech’s theatrics with their own data sovereignty rather than with fear” [5]. The commentary, framed around the idea that the Trump administration is “exposing” Big Tech’s show, urged Bern and Berlin not to be deceived [5]. This tone, aimed at the specific governments of Switzerland and Germany, shows that data sovereignty is being treated not as an abstract principle but as a substantive policy direction in European policy discourse.
2. Current Situation
In Canada, concerns over the U.S. CLOUD Act are crystallizing into a debate over securing the country’s own telecommunications sovereignty. According to a report in The Globe and Mail, the Ottawa government’s attempt to build a sovereign telecom network faces significant hurdles in terms of cost and complexity [11]. The same article reported that several European governments, worried that the CLOUD Act could compel Microsoft to disclose their public sector data to the U.S. government, are migrating public data to open-source or domestic solutions [11]. This represents a national-level institutional response, distinct in nature from the local-level opposition in France and Austria.
Political resistance to data centers is also growing within the United States. BloombergNEF analyzed that state governments in Texas, New York, and Pennsylvania are moving to delay or halt new data center projects due to concerns about electricity costs, infrastructure strain, and local impacts [2]. The report noted that Texas accounts for about 20% of the entire U.S. data center pipeline, meaning even short delays could result in billions of dollars in losses for developers [2]. The Globe and Mail reported that AI has become a hot-button political issue ahead of the November midterm elections, with growing calls for regulation from Silicon Valley to Washington, D.C. At the same time, local conflicts over the construction of data centers, which consume vast amounts of power and water, are erupting in large and small cities across the United States [15]. The Brookings Institution assessed that data centers have emerged as a key issue for the 2026 midterm elections, citing examples such as an ad by Democratic candidate Sherrod Brown in the Ohio Senate race targeting the Republican incumbent as the “face of data centers” [12].
In response, former President Trump has squarely rejected demands for regulation, labeling the opposition to AI and data centers as a “sick conspiracy theory” [17]. This stance serves as a rebuttal to both the data sovereignty discourse emerging in Europe and the backlash within the United States. European Central Bank President Christine Lagarde warned that if Europe fails to build its own AI technology and computing capacity, it could face an unprecedented risk of having its access to AI completely cut off [16]. The Chinese state-run Global Times cited this remark to highlight the risks that the “tech right” movement in the U.S. poses to Europe, reflecting a Chinese perspective that seeks to emphasize the U.S.-Europe technology dependency structure, separate from the U.S.-China conflict frame [16].
3. Key Actors and Positions
Local governments and residents in France and Austria are the primary stakeholders, protesting the land encroachment, power and water consumption, and changes to the local economic structure caused by data center construction [1][13]. Their opposition is characteristically framed as an issue of local livelihood rather than a national policy agenda.
Swiss and German policy elites, as represented by the NZZ commentary, take the position that they must build their own data infrastructure without being swayed by Big Tech’s crisis narratives [5]. This shows a trend where the European-level policy discourse is converging toward a self-strengthening argument at the individual government level.
The Canadian government is elevating the potential infringement on its data access rights due to the CLOUD Act to an issue of telecommunications sovereignty, but it is constrained by implementation costs and technical complexity [11].
The U.S. administration and Big Tech consider the expansion of data centers and investment in AI infrastructure as central to their economic and security strategy, and they are on the defensive, framing regulatory demands as “conspiracy theories” [17]. In contrast, U.S. state governments and local voters are forming a bipartisan backlash due to rising electricity costs and infrastructure burdens, revealing a clear temperature difference between the federal administration and local governments [2][12].
The European Central Bank and other European monetary and policy institutions are wary of the possibility that dependence on U.S. AI infrastructure could be used as leverage in future negotiations on issues like tariffs and digital taxes [16]. This suggests that Europe’s digital sovereignty discourse is being perceived not just as an industrial policy matter but as an issue of trade negotiating power.
4. Key Issues
The first issue is that local conflicts over the physical location of data centers and national-level concerns about sovereignty infringement due to extraterritorial jurisdiction issues like the CLOUD Act are proceeding simultaneously on different levels. The cases in France and Austria correspond to the former, while the Canadian case corresponds to the latter [1][13][11].
The second issue is the fact that even within the United States, attitudes toward data centers are divided between the federal and local levels. The federal administration views AI infrastructure expansion as a matter of national competitiveness, whereas state governments and voters are applying the brakes due to power shortages and local burdens [2][17].
The third issue is that while Europe’s response is unfolding on both the discursive level (the NZZ-style self-strengthening argument) and the practical policy level (migrating public data, adopting domestic solutions), there is a significant structural gap with U.S. Big Tech in terms of implementation capacity and financial resources [5][11][4]. This highlights how data center and cloud infrastructure have emerged as key assets in the competition over technology standards and industrial policy from a trade and economic security perspective. Within the EU, the fact that this problem is manifesting as fragmented policy responses at the individual member state level also exposes the limits of intra-regional coordination.
II. In-depth Analysis
The Threat to “Digital Sovereignty” from Big Tech and the Proliferation of National Responses: An In-depth Analysis
1. Analysis of Root Causes
At the root of this conflict lies the structural asymmetry of cloud and data center infrastructure being concentrated in the hands of a few U.S. companies. The degree to which European governments and businesses rely on the services of Microsoft, Google, and Amazon has already reached an irreplaceable level. A study by Roland Berger diagnosed a structurally widening gap between Europe and the United States in the AI infrastructure race [4]. This gap is not merely a matter of industrial competitiveness. The core issue is that this infrastructure dependence translates directly into an asymmetry in negotiating power.
The warning from ECB President Lagarde accurately pinpoints the nature of this asymmetry. She noted that “any U.S. restriction of access or change in conditions would have simultaneous knock-on effects on all sectors” [16]. She described this as “a kind of leverage that no trading partner has ever exercised over Europe” and warned that this leverage could be brought to bear in negotiations on tariffs or digital taxes [16]. In other words, Europe’s concern is not technological dependence itself, but the possibility of that dependence being repurposed as a weapon in trade negotiations. This is the point where competition over technology standards and industrial policy directly intersects with the issue of data sovereignty in the realm of trade and economic security.
The U.S. CLOUD Act entrenches this structure at the legal and institutional level. The law’s extraterritorial application provision, which allows the U.S. government to access data stored abroad by U.S. companies with a warrant, was the direct trigger for Canadian and European governments to fundamentally reconsider storing their public data on U.S. clouds [11]. The Globe and Mail reported that several European governments, concerned that the CLOUD Act could compel Microsoft to hand over their data to the U.S. government, are migrating to open-source or domestic solutions [11]. This is a classic example of how cybersecurity and digital infrastructure sovereignty are converging in the domain of emerging and non-traditional security.
2. Structural Context
At the political level, a distinctive feature of the European backlash is that it first erupted at the local government level rather than the national level. The resignation of the mayor in Étrechet, France, the local protests in Kronstorf, Austria, and the belated opposition in the Kajaani region of Finland were all triggered by conflicts over local land, power, and agricultural resources, not central government policy [1][13][8]. The structural characteristic of this issue is the pathway by which this local-level resistance converges into national policy discourse, as seen in the NZZ commentary. In the process by which micro-level backlash in local politics is transformed into a macro-level framework of national sovereignty discourse, European governments are absorbing it into the policy language of data sovereignty. This is similar to the typical pathway in EU affairs where individual member state responses spread to become intra-regional policy debates.
At the economic level, competition for physical resources like electricity and water adds real weight to the sovereignty discourse. In the United States as well, the state governments of Texas, New York, and Pennsylvania are delaying data center projects due to electricity costs and infrastructure burdens [2]. Given that Texas accounts for about 20% of the entire U.S. pipeline, these delays have a ripple effect across the industry [2]. In short, the data sovereignty debate, even within the U.S., is intertwined with the resource allocation problem of “who gets the power.” Although the European data sovereignty discourse and the American local resistance use different political languages, they share the same structural grievance that physical infrastructure encroaches on local resources.
At the security level, the situation where cloud infrastructure is incorporated as a physical target in proxy wars between nations adds a new urgency to this debate. The March 2026 incident in which three AWS data centers in Bahrain suffered structural damage from an Iranian drone retaliatory strike is a case that demonstrated data centers can be geopolitical targets, not just commercial facilities [9][10]. While this is different in nature from Europe’s data sovereignty debate, it serves as a backdrop that reinforces the shared perception that the location and control of cloud infrastructure are directly linked to national security.
3. Historical Precedents and Comparison with Similar Cases
The discourse on data sovereignty itself is not new. A commentary by EAI Professor Kim Sang-bae previously analyzed that in the U.S.-China data norm competition, “the United States, which pursues international norms guaranteeing the transnational flow of data, has been competing with China, which pursues data ‘national sovereignty’ to protect its domestic data market” [3]. The commentary pointed out that the European Union has taken a third approach in this competitive landscape, “addressing data issues at the level of ‘sovereignty as an idea’” [3]. The recent backlash from Europe and Canada is distinct from previous phases in that it marks the beginning of intervention by European and North American middle powers as an independent axis in this long-standing U.S.-China data norm competition. Whereas Europe’s past data sovereignty discussions centered on rule-making, such as the GDPR, the current phase is different in that the front line has shifted to the issue of physical infrastructure ownership and location.
A comparison with the spread of techno-fascism discourse is also valid. An EAI analysis pointed out that this discourse was “sparked not in Silicon Valley but in the intellectual circles of Latin America and Europe” [7]. The recent NZZ commentary shows the same pattern of European media critically reinterpreting American Big Tech narratives and converting them into their own policy discourse [5]. Both cases show that Europe is not a passive recipient of U.S.-origin tech discourse but an actor that reinterprets and appropriates it as a means for its own policymaking. However, as the EAI analysis noted, “the UN warns against delegating development direction to a few Big Tech companies, but its effectiveness is limited” [7]. Thus, checks at the international organization level are likely to play a similarly limited role in the current data sovereignty debate.
Canada’s attempt to build its own telecommunications network follows a trajectory similar to past nationalization and domestication efforts in strategic industries. However, as The Globe and Mail points out, this attempt faces “significant hurdles in terms of cost and complexity” [11]. This overlaps with the pattern where past domestication attempts in strategic industries like telecommunications and energy were scuttled or delayed by initial cost burdens. Because data infrastructure has a shorter technology renewal cycle and requires larger initial investment than telecommunications networks, the probability of success for domestication attempts is even more uncertain than in past cases.
4. Key Variables Shaping Future Developments
The first variable is the domestic political landscape in the United States. Former President Trump has frontally rebutted demands for regulation of AI and data centers, labeling them a “sick conspiracy” [17]. This suggests that Europe’s data sovereignty discourse is unlikely to gain cooperation at the U.S. federal government level. At the same time, data centers have emerged as a key issue in the U.S. ahead of the 2026 midterm elections [12][15]. The case of tax incentives for data centers being used as a point of attack in the Ohio Senate primary race shows the possibility that the federal government’s hardline external stance and the backlash at the state and local levels could unfold separately [12]. If this divergence continues, the response strategies of Europe and Canada may also shift toward seeking individual negotiation channels with state and local actors rather than the U.S. federal government.
The second variable is whether institutional convergence occurs at the EU level. Currently, the backlash is unfolding sporadically at the individual country and regional levels in France, Austria, Finland, Switzerland, and Germany [1][13][8][5]. Whether these individual responses converge into a unified cloud sovereignty policy at the European Commission level or disperse into divergent responses by member states will determine Europe’s future negotiating power. If a unified response is formed, the practical capacity to address the negotiating leverage problem warned of by President Lagarde will increase. However, a dispersed response is highly likely to weaken the negotiating power of individual member states.
The third variable is whether middle power solidarity materializes. The “coalition of middle powers” concept proposed by Canada’s Mark Carney is an attempt to address the data sovereignty issue outside the U.S.-China bipolar framework [18]. However, it is uncertain whether this concept will move beyond symbolic diplomatic issues like a trade ban on products from Palestinian settlements to practical policy coordination in the concrete trade and technology domains of cloud infrastructure and data norms. If this solidarity materializes, it could open up opportunities for middle powers, including South Korea, to participate in negotiating platforms related to data sovereignty. For now, however, it remains at the level of discursive consensus.
The fourth variable is the speed at which the physical security risks of cloud infrastructure are reassessed. Following the attack on data centers in the Middle East, some in the U.S. policy community have begun to reconsider the concentration of frontier model training infrastructure in high-risk regions [10]. If this reassessment leads to a dual-track strategy that separates high-risk, critical workloads from low-risk, general-purpose workloads, it could lend weight to the logic of Europe and Canada demanding that critical infrastructure be hosted within their own borders. Conversely, if U.S. Big Tech responds to geopolitical risks by further concentrating infrastructure in the U.S. mainland and a few allied countries, the domestication attempts by Europe and Canada could face a paradoxical situation where their very access is restricted.
3 credits are required from here
The body beyond the scenario analysis is available with credits.
Sign in to continue reading*This text is an AI translation of an original written in Korean. Some translations or nuances may be inaccurate.
This report is an in-depth analysis planned by an EAI researcher, grounded in sophisticated AI-assisted research, and finalized by the EAI researcher.