The AI Automation Risk To Digital Sovereignty

Introduction

The rapid acceleration of artificial intelligence automation is fundamentally reshaping the concept of digital sovereignty, forcing nations to confront urgent questions about who controls the data, infrastructure and algorithms that increasingly govern modern economies and societies. As AI systems become embedded in everything from public administration and healthcare to national defense and financial markets, the ability of a state to exercise meaningful authority over its own digital future has become one of the defining geopolitical challenges of the 21st century.

The Evolving Meaning of Digital Sovereignty

Digital sovereignty refers to a country or organization’s ability to control its own digital data, infrastructure and technology without external interference. Historically, the concept was associated with data protection regulations and internet governance, but the emergence of AI automation has expanded its scope dramatically. It now encompasses not only data flows but also the computational infrastructure that processes them and the talent pipelines that sustain the entire ecosystem. As the Tony Blair Institute for Global Change has argued, sovereignty in the age of AI should not be understood as independence from all others, but rather as the ability to act strategically –  with agency and choice – in a world that is irreversibly interdependent. The concept broadly rests on two main pillars. Data sovereignty concerns how much control an entity has over the data it uses and produces and technological sovereignty, which concerns the degree of control over the digital technologies it relies upon (there are also two further related pillars of operational sovereignty and assurance sovereignty). AI automation has intensified the stakes on both fronts. The training of large language models and other AI systems requires vast quantities of sensitive data, meaning that nations without robust governance frameworks risk seeing their citizens’ information absorbed into foreign-controlled systems. At the same time, the sheer capital intensity of AI development – requiring billions of dollars in specialized hardware, energy and engineering talent – means that frontier AI capability is overwhelmingly concentrated in the United States and China, which together control more than 90 percent of global AI data-centre capacity…

The Concentration of AI Power

The structural dynamics of AI automation naturally concentrate power in the hands of a small number of actors. The United States currently hosts approximately 75 percent of the world’s total AI compute capacity, compared to about 15 percent in China and roughly 10 percent distributed elsewhere, mostly in Europe. Only 32 countries worldwide host AI-specific data centres, leaving around 160 nations entirely dependent on foreign infrastructure for their AI needs. In the European cloud market, local providers’ combined share fell from 29 percent to 15 percent between 2017 and 2024, while three US-based hyper-scalers now account for about 70 percent of demand.

Nations that cannot access or control the computational resources driving AI automation become structurally dependent on those that can

This concentration has profound implications for sovereignty. Nations that cannot access or control the computational resources driving AI automation become structurally dependent on those that can. As the World Economic Forum has noted, cross-border data flows that once seemed routine now face stricter oversight or outright restrictions under the banner of digital sovereignty, as the data AI systems rely on has turned into a strategic asset. The politicization of data is a striking feature of the current landscape. Governments increasingly view datasets not as neutral commodities but as instruments of national power that must be carefully governed.​

Europe’s Regulatory and Industrial Response

The European Union has mounted the most ambitious regulatory response to the sovereignty challenges posed by AI automation. The EU’s AI Act, which entered into force on 1 August 2024 and will be phased in gradually until full application by 2 August 2027, creates harmonized conditions for AI market access across the bloc while ensuring safety and fundamental rights protection. It classifies AI applications by risk level, banning practices such as harmful AI-based manipulation, social scoring and certain forms of biometric surveillance, while imposing transparency and compliance obligations on high-risk systems. Beyond regulation, the European Commission launched the AI Continent Action Plan in April 2025, a €200 billion strategy to create a sovereign, pan-European AI ecosystem. The plan aims to triple the EU’s AI compute capacity by 2027 through so-called AI Factories and forthcoming Gigafactories, while a new Data Union Strategy seeks to unlock sector-specific datasets for European innovators. In November 2025, EU Member States signed the Berlin Declaration for European Digital Sovereignty, which highlighted the need to mitigate digital dependencies and advance the EU’s technological capabilities. However, critics noted it lacked an explicit commitment to fundamental rights enforcement.

France has emerged as Europe’s most assertive player in sovereign AI infrastructure

France has emerged as Europe’s most assertive player in sovereign AI infrastructure. The country has committed over €109 billion in AI-related investment through 2030, anchored by the Paris-based startup Mistral AI, which develops open-weight large language models designed to rival American counterparts while remaining fully compliant with European regulations. At VivaTech 2025, French President Emmanuel Macron appeared alongside Nvidia’s Jensen Huang and Mistral CEO Arthur Mensch to affirm a shared commitment to building a sovereign European AI based on a local value chain from chips and data to models. The launch of Mistral Compute, a European AI computing infrastructure developed with Nvidia, represents a tangible effort to give Europe control over its own technology and data, offering an alternative to dependence on US hyperscalers.

China’s Drive for AI Self-Reliance

China has pursued a markedly different path toward AI sovereignty, one shaped by state direction, massive public investment and an explicit goal of reducing dependence on Western technology.

Under President Xi Jinping, Beijing has made “independent and controllable” AI a key national objective, seeking self-reliance at every level of the technology stack from hardware to algorithms. China’s AI strategy rests on three pillars.

• Building a self-sufficient ecosystem to reduce foreign dependence on chips and algorithms
• Embedding AI across the economy and defense
• Exporting governance models worldwide through initiatives like the Global AI Governance Initiative.

The urgency of this drive has been intensified by US semiconductor export controls, which have limited China’s access to the most advanced AI chips. In response, Beijing has reportedly mandated that all state-funded data centres under construction must use domestically developed AI chips, excluding components from American companies like Nvidia and Intel. In 2025 alone, government funding accounted for approximately 400 billion yuan of the nation’s projected 600 to 700 billion yuan in AI capital expenditure. Companies like Baidu have begun training new AI models using in-house Kunlun chips, while Cambricon Technologies has reported a fourteen-fold revenue increase driven by surging orders for its domestic AI accelerators. China’s approach thus treats AI sovereignty not merely as a matter of economic competitiveness but as a strategic imperative inseparable from national security and military readiness.

China’s approach thus treats AI sovereignty not merely as a matter of economic competitiveness but as a strategic imperative inseparable from national security and military readiness.

The Sovereign Cloud as a New Arena

One of the most tangible manifestations of the tension between AI automation and digital sovereignty is the rise of sovereign cloud computing. These are cloud environments designed to guarantee that sensitive information remains within the jurisdiction of the host country, protected from geopolitical conflicts and global network outages. Sovereign clouds are built around three key dimensions.

1. Infrastructure sovereignty, ensuring locally controlled hardware
2. Operational sovereignty, maintaining trusted personnel and processes
3. Software sovereignty, guaranteeing the ability to run applications without excessive dependence on foreign suppliers.

The demand for sovereign cloud solutions has grown rapidly as AI workloads increasingly require processing sensitive government and enterprise data. NATO, for example, signed a multimillion-dollar contract with Google Cloud in November 2025 to deliver an air-gapped sovereign cloud environment capable of running AI models and analytics on classified data while preserving strict data residency and operational control. In France, Capgemini and Orange jointly created Bleu, a company offering Microsoft-based cloud services that meet French sovereignty standards, targeting critical infrastructure operators and public institutions in regulated industries. However, as analysts at the Center for Strategic and International Studies have warned, sovereign clouds offer greater control but do not necessarily provide greater technical security. The higher costs, slower growth and reduced innovation they bring can make the economies that rely on them less competitive.

The higher costs, slower growth and reduced innovation they bring can make the economies that rely on them less competitive

Implications for Developing Nations

The impact of AI automation on digital sovereignty is especially acute for developing countries, which face the risk of being locked out of the AI value chain altogether. While AI systems promise breakthroughs in health, education climate resilience and other major domains, they also risk deepening digital dependency, enabling unchecked surveillance, and accelerating job displacement in nations that lack the infrastructure and institutional capacity to govern these technologies effectively. Developing countries experience a lower degree of exposure to AI compared to high-income countries, partly because of a predominance of manual labor and partly because of inadequate access to essential infrastructure such as electricity and reliable internet.

The emerging threat is particularly stark in digital services

The emerging threat is particularly stark in digital services. Many developing countries had bet on business process outsourcing, call centres, and data labelling as pathways to economic development, yet AI automation is now capable of performing precisely these tasks at lower cost. If AI systems can function as what some experts call “drop-in remote workers” by the end of the decade, they could strip away several early rungs of the export-driven development ladder, denying developing nations the competitive advantages they once held in labor arbitrage. For African economies, the stakes are especially high, as the risk of marginalization in AI governance remains significant unless representation and coordination mechanisms are dramatically strengthened…

Reshaping of Supply Chains

AI automation is also reshaping global supply chains in ways that have direct consequences for digital sovereignty. The convergence of geopolitical tensions, tariff regimes and AI-driven productivity gains has accelerated re-shoring and near-shoring trends, as companies seek to reduce their dependence on distant and potentially adversarial suppliers. In the United States, 2025 tariff rates reached 18.6 percent – the highest since 1933 – prompting 18 percent of manufacturers to shift production domestically within six months, with semiconductor and electric vehicle battery plants leading the charge. AI-powered analytics are playing a crucial role in enabling these transitions, helping businesses identify cost-effective re-shoring opportunities, mitigate risks and optimize supplier networks. The adoption of Industry 4.0 technologies, including AI-driven demand forecasting, robotics, and digital twins, has become an enabler for supply chain realignment, allowing organizations to build more resilient and responsive operations closer to home. This dynamic reinforces the link between AI automation and sovereignty. Nations that can deploy AI to strengthen their domestic industrial base gain strategic autonomy, while those that cannot risk further marginalization in global value chains.

Nations that can deploy AI to strengthen their domestic industrial base gain strategic autonomy, while those that cannot risk further marginalization in global value chains

Toward a Distributed Architecture

A growing body of evidence suggests that the next phase of AI development may help reconcile the tension between sovereignty and competitiveness rather than deepening it. As AI moves from monolithic large language models toward  “agentic AI” –  networks of specialized agents that collaborate and act in real time at the edge – the architecture of AI is becoming inherently more distributed. Many of the tasks that create the greatest added value rely on local, sensitive data, meaning that fine-tuning and inference increasingly happen in environments the data owner controls, such as an enterprise data centre, a hospital campus, a factory floor or a trading desk.​ This shift toward distributed, hybrid architectures means that different countries and regions can play distinct but interoperable roles in the AI value chain. Hyperscale facilities may continue to concentrate the training of very large models, but regional centres can handle fine-tuning with proprietary data, while edge nodes embedded in factories, vehicles or telecommunications exchanges can perform real-time inference. Because contribution is possible at every layer, countries can specialize according to their comparative advantages, whether that is abundant renewable energy, advanced manufacturing data, a strong healthcare system, or robust regulatory frameworks for sensitive information. In this emerging model, competitiveness and sovereignty become complementary rather than conflicting objectives.​

Because contribution is possible at every layer, countries can specialize according to their comparative advantages…

Conclusion

Full self-sufficiency in AI is neither feasible nor desirable for most nations

The intersection of AI automation and digital sovereignty will remain one of the most consequential policy arenas of the coming decade. The choices that governments make today about their AI infrastructure, regulatory frameworks, talent pipelines and international partnerships will shape whether they retain meaningful agency over their digital futures or cede that agency to a handful of foreign corporations and states. Full self-sufficiency in AI is neither feasible nor desirable for most nations, but neither is passive dependence on systems developed and governed elsewhere. The path forward lies in what might be called strategic interdependence: building domestic strengths where they matter most, securing reliable access to frontier capabilities through carefully negotiated partnerships, and investing in the institutions and governance structures needed to ensure that AI automation serves national interests rather than undermining them.​As Goldman Sachs has observed, geopolitical swing states and blocs will shape the future of AI through their economic and regulatory power, their differentiated technology ecosystems, their control over critical supply chain chokepoints, as well as their capability and will to implement clear national AI strategies. The question is no longer whether AI automation will transform the foundations of sovereignty, but whether nations can summon the foresight and co-ordination to ensure that transformation strengthens rather than erodes their capacity for self-determination.​

References:

S. Soh, “Digital Sovereignty in the Age of AI,” IT Connection, May 2025. https://itcblogs.currentanalysis.com/2025/05/07/digital-sovereignty-in-the-age-of-ai/​

William Fry, “Europe’s AI Ambitions: Inside the EU’s €200 Billion Digital Sovereignty Plan,” April 2025. https://www.williamfry.com/knowledge/europes-ai-ambitions-inside-the-eus-e200-billion-digital-sovereignty-plan/​

World Economic Forum, “AI Geopolitics and Data in the Era of Technological Rivalry,” January 2026. https://www.weforum.org/stories/2025/07/ai-geopolitics-data-centres-technological-rivalry/​

Tony Blair Institute for Global Change, “Sovereignty in the Age of AI: Strategic Choices, Structural Dependencies,” September 2025. https://institute.global/insights/tech-and-digitalisation/sovereignty-in-the-age-of-ai-strategic-choices-structural-dependencies​

European Commission, “AI Act: Shaping Europe’s Digital Future,” January 2026. https://digital-strategy.ec.europa.eu/en/policies/regulatory-framework-ai​

Goldman Sachs, “The Generative World Order: AI, Geopolitics, and Power,” December 2023. https://www.goldmansachs.com/insights/articles/the-generative-world-order-ai-geopolitics-and-power​

BNP Paribas, “AI, Digital Sovereignty, Cybersecurity, Health: Four Key Trends at VIVA Technology 2025,” June 2025. https://group.bnpparibas/en/news/ai-digital-sovereignty-cybersecurity-health-four-key-trends-at-viva-technology-2025​

Verfassungsblog, “EU’s Digital Sovereignty and the Rights-Based Imperative,” December 2025. https://verfassungsblog.de/digital-sovereignty-and-the-rights/​

IZA, “Artificial Intelligence and the Future of Work: Evidence and Policy,” 2025. https://docs.iza.org/pp216.pdf​

Ashley Dudarenok, “China AI Strategy: Policy, Regulation & Global Impact in 2025,” January 2026. https://ashleydudarenok.com/china-ai-strategy/​

Prism Media, “NATO Buys Google Cloud Sovereign Platform to Run Classified AI,” November 2025. https://www.prismedia.ai/news/nato-buys-google-cloud-sovereign-platform-to-run-classified-ai​

AI Frontiers, “AI Could Undermine Emerging Economies,” February 2026. https://ai-frontiers.org/articles/ai-could-undermine-emerging-economies​

Keyss Inc., “China’s AI Chip Mandate 2025–2026,” January 2026. https://keyssinc.com/china-ai-chips-2025/​

T20 South Africa, “From Digital Dependence to Digital Sovereignty,” September 2025. https://t20southafrica.org/commentaries/from-digital-dependence-to-digital-sovereignty-south-africas-g20-opportunity-in-the-age-of-ai/​

MERICS, “MERICS Report: China’s AI Stack,” July 2025. https://merics.org/sites/default/files/2025-07/MERICS%20Report-AI_Stack_final.pdf​

World Economic Forum, “How AI Can Balance Competitiveness and Digital Sovereignty,” January 2026. https://www.weforum.org/stories/2026/01/ai-s-distributed-future-a-new-path-to-competitiveness-and-digital-sovereignty/​

China US Focus, “China’s Strive for Self-Reliance in Advanced Technology,” October 2025. https://www.chinausfocus.com/peace-security/chinas-strive-for-self-reliance-in-advanced-technology​

BCG, “Keys to a Successful Sovereign Cloud,” June 2025. https://www.bcg.com/publications/2025/sovereign-clouds-reshaping-national-data-security​

CSIS, “Sovereign Cloud–Sovereign AI Conundrum: Policy Actions to Achieve Prosperity and Security,” April 2025. https://www.csis.org/analysis/sovereign-cloud-sovereign-ai-conundrum-policy-actions-achieve-prosperity-and-security​

ISCEA, “AI & Reshoring – The Future of Supply Chain Transformation,” June 2025. https://www.iscea.org/post/ai-reshoring-the-future-of-supply-chain-transformation​

Introl, “France’s AI Sovereignty Push: Infrastructure Behind the European AI,” September 2025. https://introl.com/blog/france-ai-sovereignty-mistral-sovereign-cloud-2025​

AInvest, “Supply Chain 2.0: AI-Driven Reshoring and the Reshaping of U.S. Equities in 2026,” December 2025. https://www.ainvest.com/news/supply-chain-2-0-ai-driven-reshoring-reshaping-equities-2026-2601/​

CPSCP, “Reshoring and Nearshoring Trends Impacting US UK Manufacturing 2025,” November 2025. https://cpscp.org/reshoring-and-nearshoring-trends-impacting-us-uk-manufacturing-2025/​

GDPR Local, “Data Sovereignty and Digital Sovereignty,” August 2025. https://gdprlocal.com/digital-sovereignty/​

Apizee, “What is Digital Sovereignty? Digital Assets and Governance,” February 2026. https://www.apizee.com/digital-sovereignty.php