Francesca Bria is the initiator of the EuroStack Project, an innovation economist and a leading expert in digital policy. She is a professor at University College London’s Institute for Innovation and Public Purpose, an adviser to the European Commission and the former chief technology officer of the city of Barcelona. She currently serves as president of the Italian Innovation Agency, Emilia-Romagna Region.
Editor’s Note: Noema is committed to hosting meaningful intellectual debate. This piece is in conversation with another, written by Benjamin Bratton, the Berggruen Institute’s Antikythera program director. Read it here: “Is European AI A Lost Cause? Not Necessarily.“
The Cartography Of Digital Power
Geopolitical power once flowed through armies and treaties, but today it courses through silicon wafers, server farms and algorithmic systems. These invisible digital infrastructures and architectures shape every aspect of modern life. “The Stack” — interlocking layers of hardware, software, networks and data — has become the operating system of modern political and economic power.
The global race to control the Stack defines the emerging world order. The United States consolidates its dominance through initiatives like Stargate, which fuses AI development directly to proprietary chips and hyperscale data centers, creating insurmountable barriers to competition. China advances through systematic industrial policy and its Digital Silk Road, achieving unprecedented integration from chip design to AI deployment across Asia and beyond. These are deliberate strategies of technological imperialism.
Europe occupies a paradoxical position: a regulatory leader but infrastructurally dependent. We Europeans have set global standards through GDPR and the AI Act. Our research institutions remain world-class. Yet just 4% of global cloud infrastructure is European-owned. European governments, businesses and citizens depend entirely on systems controlled by Amazon, Microsoft and Google — companies subject to the U.S. CLOUD Act’s extraterritorial surveillance requirements. When we use “our” digital services, we’re actually using American infrastructure governed by American law for American interests.
This dependency isn’t abstract — it’s existential. In the 21st century, those who control digital infrastructure control the conditions of possibility for democracy itself. Europe faces a choice: build sovereign technological capacity or accept digital colonization.
The Charges Against European Tech Discourse
The attempt to blame Europe’s digital paralysis on its critical intellectuals — those who resist Silicon Valley accelerationism, crypto hyper-libertarianism and techno-authoritarianism — is profoundly misdirected. At a Venice Architecture Biennale event I organized, “Archipelagos of Possible Futures,” Benjamin Bratton leveled four charges against Europe’s approach to AI and digital sovereignty. He argued, first, that Europe follows a “regulate first, build later (maybe)” strategy that breeds dependency rather than sovereignty; second, that its tech critics are “technophobic public intellectuals” who only say why not to build; third, that this critical culture produces “analysis paralysis” that blocks the very innovation it seeks; and fourth, that environmental and social concerns are manipulated to defend intellectual status rather than confront real technological challenges.
Each charge misses the mark. Europe’s technological predicament is not the result of excessive critique but of three decades spent dismantling the very capacities needed for technological sovereignty. The real choice we face is not between criticism and construction, but between authoritarian technological models and democratic alternatives. And building such alternatives demands that we understand how power has historically operated through technology. To dismiss, as Bratton does, thinkers like Evgeny Morozov, Kate Crawford and Marina Otero, is to embrace precisely the techno-accelerationist logic I believe must be confronted today.
Understanding the material realities of AI infrastructure — its environmental costs, labor dependencies and power concentrations — isn’t “fearmongering” but a prerequisite for sustainable, democratic development. And Europe’s emerging independent technological ecosystem demonstrates that democratic construction is already happening when we create the right conditions.
Most fundamentally, dismissing regulation as a European pathology — in Bratton’s phrasing, “the EU has AI regulation but not much AI to regulate” — reveals profound superficiality. Regulation is not the problem. The problem is a lack of enforcement and the absence of an industrial policy at scale. Indeed, if European regulation were so ineffective, why did the Trump administration go so far as to threaten bans and sanctions against European regulators who dared to implement digital laws? Precisely because effective regulation is seen in Washington as a direct obstacle to U.S. technological supremacy.
The Roots Of Europe’s Technological Predicament
The roots of Europe’s predicament are to be found, first, in the stranglehold of neoliberal orthodoxy on European economic thinking. This caused decades of austerity and made Europe the poster child of hyper-globalization: free trade without economic statecraft, bans on state aid, trickle-down economics dogmas, no long-term public investment and the systematic rejection of industrial policy. All of this was gospel preached by the U.S., even as Washington protected its own national security interests and subsidized Silicon Valley. Europe, in other words, internalized the ideology of market neutrality while others practiced strategic capitalism.
American technological dominance wasn’t born from free markets but from massive state intervention — DARPA, NASA and the National Science Foundation provided the patient capital and guaranteed markets that made Silicon Valley possible. As Linda Weiss’s work on the national security state and Mariana Mazzucato’s research on the entrepreneurial state have documented, innovation ecosystems do not emerge spontaneously — they are structured through public direction. Every core technology in the iPhone — the internet, GPS, touchscreens, voice recognition — emerged from decades of public research funding. The Pentagon’s procurement budgets functioned as venture capital at a continental scale.
Meanwhile, Europe internalized market fundamentalism more thoroughly than its own inventors intended. The Stability and Growth Pact threatened strategic industrial investment as a violation of fiscal discipline. Competition policy, instead of preventing Big Tech dominance via bold antitrust action, prevented the formation of European champions while American firms achieved monopolistic scale. Europeans were lectured that industrial policy violated market principles by the very Americans who systematically practiced it.
“In the 21st century, those who control digital infrastructure control the conditions of possibility for democracy itself.”
The CLOUD Act handed U.S. agencies jurisdiction over European data, digital trade agreements prevented data localization and intellectual property regimes ensured that value extraction flowed westward across the Atlantic. Rather than using its regulatory power to block predatory surveillance models and turn data sovereignty into a competitive advantage, Europe tolerated a system now weaponized by far-right tech oligarchs to spread disinformation, extremism and fake news.
This isn’t a cultural failure — it’s a political one. I’ve seen it first-hand as CTO of the city of Barcelona, president of the Italian Innovation Fund and coordinator of major European research projects. When European startups succeed, they turn to American venture capital and often relocate to Silicon Valley. When our researchers make breakthroughs, they are hired away by U.S. companies offering vastly higher pay. When European cities try to assert digital sovereignty, they face lawsuits bankrolled by American tech giants and diplomatic pressure from Washington. Powerful forces, in other words, are arrayed against anyone who tries to change course. The result is a steady drain of talent, capital and sovereignty.
Who Controls The Stack? The New Techno-Economic Warfare
Dismissing European concerns about technological sovereignty as “Cold War thinking” misreads reality. This isn’t ideological competition — it’s economic warfare where technology is a national weapon built through political and industrial choices.
Control concentrates at every layer of the stack. In materials, China processes 90% of global rare earths. So when it restricts gallium, germanium and graphite exports, it strikes directly at European and other green energy transitions.
At the chip layer, Taiwan’s TSMC commands 64% of global foundry capacity, Samsung another 12%. Europe has fallen to 8%, despite ASML’s lithography monopoly. The Trump administration’s 10% equity stake in Intel signals how far Washington wants to go. Nvidia and AMD agreed to hand over 15% of their AI chip revenues to the U.S. government just so they could keep selling into China. U.S. export controls don’t just constrain China — they dictate what European firms can sell and what researchers can access. The Dutch licensing restrictions on ASML, one of the world’s leading suppliers of equipment essential for making computer chips, show how American regulations reverberate across Europe’s industrial core.
But export controls reveal their limits. China’s DeepSeek achieved competitive AI performance at a fraction of typical costs. In response, some leaders in Silicon Valley and Washington called for even tighter restrictions on AI chips and infrastructure, pushing China toward self-sufficiency while fragmenting the global tech stack further.
“Innovation ecosystems do not emerge spontaneously — they are structured through public direction.”
At the cloud and AI layers, U.S. hyperscalers dominate. The CLOUD Act grants Washington extraterritorial reach over any data touching U.S. companies — even when stored in Europe. Nearly all foundation models answer to Silicon Valley.
But AI dominance now comes packaged with ideology. Trump’s executive orders mandate AI systems “free from ideological bias,” banning “woke AI” in federal procurement while defining diversity and equity as distortions that “sacrifice truthfulness.” The U.S. AI Action Plan exports “its full AI technology stack—hardware, models, software, applications, and standards—to all countries willing to join America’s AI alliance.” This isn’t just chips — it’s the entire stack, with American values and control baked in at every layer.
Trump made it explicit: “substantial” tariffs against any country regulating U.S. tech firms. Thus, Europe cannot set rules in its own market without facing economic punishment. One executive order in Washington — not Brussels or Berlin — could cut access to critical systems running our industries, hospitals and elections. That’s not a trade deficit — it’s a sovereignty deficit.
Whoever controls AI infrastructure — compute, models, data and cloud — will shape the economic and political order of the 21st century. The U.S. and China understand this and are mobilizing every instrument of statecraft to secure supremacy. Europe must understand it too.
The Material Realities Of AI Dominance
AI isn’t magic or ethereal — it’s brutally material, requiring specific configurations of energy, water, land and capital that define the political geography of digital sovereignty. Understanding these material flows reveals where power concentrates and where intervention becomes possible.
The numbers are staggering. Training frontier models consumes enormous computational resources: GPT-4’s training required electricity equivalent to the annual consumption of thousands of American homes. Google’s emissions surged nearly 50% in the past five years, driven mostly by AI computation. By 2030, data centers are projected to use at least 3% of global electricity, with AI workloads accounting for the majority. Already in Ireland, data centers consume more than a fifth of the national electricity grid, which some projections predict will rise to a third by 2030. Similar crises emerge in Frankfurt, Amsterdam, London — wherever cloud infrastructure concentrates.
Follow the money: BlackRock deploying hundreds of billions of dollars for new data center build-outs, Saudi sovereign wealth funds recycling fossil fuel profits into AI ventures, Emirati sovereign funds seeking technological hedges against energy transition.
The human costs reveal similar political economy dynamics. Kenyan workers earning minimal wages label content to train ChatGPT’s systems. Congolese children mine cobalt for data center batteries. Filipino moderators develop trauma from endless exposure to violence and abuse so AI can appear “safe.” These systems depend on hidden armies of exploited workers performing “ghost work” that makes AI seem magical.
“There is no contradiction between embracing the green agenda and having a strong AI-focused industrial policy.”
Europe’s advantage lies in treating constraints as design opportunities. Its renewable leadership — Germany’s 62% renewable mix, Spain’s solar surge, Denmark’s wind-powered grid — already provides the foundation for sustainable AI. Labor protections curb the exploitation rife in U.S. and Chinese supply chains, while environmental commitments push innovation beyond extractive models. DeepSeek proves that high-quality models can be built with less compute, fewer cutting-edge chips and lower resource use through open source and better engineering. Europe doesn’t need mega-infrastructures financed by fossil wealth — it needs models tailored to its own industries and societies.
AI’s soaring energy demands are already driving the industry back to nuclear: Amazon, Google and Microsoft are investing billions in small modular reactors. When Peter Thiel conflates Greta Thunberg with the Antichrist for defending climate action, the stakes are clear: The renewable transition is cast as a threat to innovation. Yet Europe’s renewable base is not a weakness — it is the very foundation of sustainable AI.
There is no contradiction between embracing the green agenda and having a strong AI-focused industrial policy. As carbon costs rise and resources tighten, fossil-fuelled AI will become increasingly fragile. By powering data centers with clean energy, limiting water use and pricing carbon at its real cost, Europe can turn constraint into strength.
Silicon Valley’s Turn To Techno-Nationalism
What Silicon Valley presents as neutral technological progress is increasingly revealing itself as an authoritarian political project. Trump’s second administration has accelerated this transformation with breathtaking speed. The Pentagon now directly commissions tech executives into military ranks through programs like Detachment 201. Palantir’s $10 billion U.S. Army contract makes its surveillance systems the de facto operating system of the modern military, integrating battlefield intelligence with domestic data. Anduril’s autonomous weapons factories mass-produce AI-powered drones while its executives rotate into senior Pentagon positions.
The architects of this system no longer hide their vision. Palantir CEO Alex Karp’s manifesto “The Technological Republic” articulates “patriotic tech” as a kind of fusion of Silicon Valley libertarianism with authoritarian nationalism. This ideology, rooted in anti-democratic philosophies, casts technological supremacy as a civilizational imperative.
Alex Karp presents Palantir as a bulwark against “American decay,” Elon Musk unilaterally restricts Ukrainian access to Starlink based on his own political whims and Peter Thiel directs ideological allies into government, channeling U.S. venture capital and defense money into his causes. Every major AI lab now depends on people and institutions that are opposed to democratic governance. What is emerging is not a planetary commons but a new tech-military complex financed by capital aligned with authoritarian ideologies and legitimized through patriotic rhetoric.
When critics dismiss concerns about AI bias, surveillance capitalism and platform monopolization as ideological extremism, they reveal their own allegiance to this kind of oligarchic control. To brand all this critique as “Lysenkoism” or “woke Marxism” while ignoring the real risks of AI capture echoes Trump’s new McCarthyism against imagined threats.
European Alternatives Beyond Silicon Valley
The EuroStack builds on Europe’s independent tech, research and industrial ecosystem — the foundation for democratic digital sovereignty linking demand and supply. Some argue Europe should focus on AI “diffusion” rather than infrastructure, treating computation as a “planetary common” to be accessed. Bratton accepts this framework: Europe as consumer, not builder.
But this misses the point. Infrastructure sovereignty is political agency. Control determines whether technology serves social, economic and ecological goals — or whether those goals are reshaped by Big Tech’s imperatives.
“Diffusion” without sovereignty inverts the relationship. Instead of technology serving democratically chosen ends, societies bend to platforms built elsewhere. During COVID, European governments had to follow digital protocols dictated by U.S. firms. The argument boils down to this: Everyone uses AI, but Silicon Valley or Beijing decides what AI exists, which values it encodes and which interests it serves.
Infrastructure is where power is encoded and political choices become technical constraints. Sovereignty means aligning technology with Europe’s social model, climate goals and democratic values. Without this, “diffusion” is nothing more than the efficient distribution of dependency.
For over 15 years, I’ve worked with cities and nations in Europe to turn critique into practice. In Barcelona, Mayor Ada Colau and I rejected the “smart city” model pushed by Big Tech to reimagine how technology could serve democracy. We rewrote procurement to prioritize open source and data sovereignty, launched Decidim — where 70% of city decisions came from citizen deliberation — built systems for digital rights and cryptographic data control and implemented the “Public Money? Public Code!” policy. Cities from Amsterdam to New York followed Barcelona’s example — proof that technology could serve participation over extraction.
The lessons were clear: Sovereignty starts with democratic control of infrastructure, open source enables innovation and citizens demand agency. The fiercest resistance comes not from critics but from incumbents and institutional inertia. That same logic guided my work as the president of Italy’s National Innovation Fund, where state-backed venture capital built deep-tech capacity. Europe’s strengths in bio and healthcare tech, space exploration, quantum computing and advanced manufacturing are proof that deliberate industrial strategy works.
“Infrastructure sovereignty is political agency. Control determines whether technology serves social, economic and ecological goals.”
EuroStack grows from this ground. It isn’t abstract — it’s backed by over 200 European businesses, officially endorsed by France and Germany in national strategies. The infrastructure is taking shape. Schwarz Group’s STACKIT delivers sovereign enterprise cloud from European data centers, giving businesses GDPR-compliant alternatives. OVH, Europe’s largest independent cloud provider, challenges AWS and Azure on European terms. Proton secures communications under Swiss privacy law, showing Europe can compete on security. Ionos’s Nextcloud offers true data sovereignty for collaborative work. EuroHPC pools resources into a continental supercomputing network, giving scientists, startups and industries access to world-class public compute.
On the AI front, Europe shows what deliberate strategy can achieve. Mistral, backed by €1.3B from ASML, patient public capital and French research networks, has become Europe’s leading AI startup. OpenEuroLLM develops models on European data under EU law. Switzerland’s Apertus trains open multilingual models across 1,800+ languages. The Digital Commons initiative builds open-source infrastructure, while Europe invests heavily in RISC-V adoption for computational sovereignty.
But scale matters. Mistral’s valuation is a fraction of OpenAI’s, and its models still rely on Nvidia chips and U.S. cloud infrastructure. Building alternatives means little if European firms continue defaulting to ChatGPT. The real test isn’t technical capability — it’s whether Europe can secure adoption at scale and turn these building blocks into a coherent ecosystem before dependencies become irreversible.
To drive adoption and incentivize European alternatives through strategic procurement or “Buy European” measures, leaders must stop falling for sovereignty-washing — parroting Big Tech’s AI narratives while undermining real autonomy. The U.S.-U.K. Tech Prosperity Act isn’t a path to prosperity — it leads to digital dependency that risks binding Europe tighter to American infrastructure. Every “sovereign AI” deal with NVIDIA, Google, Amazon, OpenAI or Palantir requires hard questions: Who controls the hardware? Which security laws apply? Can vendors resist foreign data demands and export controls? Who captures the value — society or monopolies?
The U.S. and China show that world-class platforms are built on decades of patient institutional funding and structural control — not venture capital alone. Europe’s distinction must lie in its values and political imagination. Silicon Valley optimizes for extraction, Beijing for control. Europe must optimize for empowerment — distributing agency rather than concentrating it.
Sovereignty Through International Digital Cooperation
The EuroStack cannot succeed in isolation. Sovereignty is not autarky. It is strategic independence: shaping technology trajectories, investing long-term, enforcing democratic accountability and building partnerships on shared principles rather than new dependencies.
Unexpected partners are emerging. India’s Digital Public Infrastructure and Brazil’s PIX payment system show how governments can build platforms that serve hundreds of millions without corporate intermediation — though privacy and rights concerns remain in Aadhaar’s implementation. Japan, South Korea and Taiwan offer manufacturing and semiconductor partnerships. Australia, Africa and Latin America bring critical mineral resources and opportunities for collaborative digital commons infrastructure beyond extractive models. These alliances must be substantive — diversifying dependencies, co-developing technologies and setting global standards.
Europe’s opportunity is leading a coalition for digital independence — public-interest AI, sovereign infrastructure, data governance, sustainable supply chains, environmental accountability. Unlike Washington or Beijing, Europe can offer technological partnership without imperial ambition: mutual benefit over dependency, open standards over lock-in, shared digital commons over monopoly control, strong governance and interoperability over surveillance capitalism or state control.
Digital Sovereignty As Democratic Power
This perspective rejects false binaries between Silicon Valley and stagnation, between digital colonialism and analog irrelevance, between planetary evolution and Luddite retreat. These illusions serve those profiting from the status quo by making alternatives unthinkable.
Europe need not choose between innovation and regulation, efficiency and equity, capability and values. The real choice is between democratic and authoritarian technology, sustainable and extractive infrastructure, distributed and concentrated power.
When European hospitals deploy AI diagnostics under strict accountability, they prove that efficiency doesn’t require abandoning oversight. When climate scientists train models on shared data protected by GDPR, they prove that ethics powers innovation. When institutions build open-source AI respecting both privacy and creators’ rights, they prove that democracy strengthens capability.
“Europe’s infrastructural future must encode democracy, sustainability and human dignity.”
The EuroStack isn’t nationalism or autarky. It demonstrates that democratic societies can shape technology, that public interest can outweigh private extraction, that human flourishing matters more than shareholder value and that sovereignty and cooperation reinforce each other.
We can accept permanent dependency, hoping foreign powers govern global infrastructures in our interest. Or we can build democratic alternatives rooted in Europe’s climate commitments, labor protections and social diversity.
Infrastructure encodes power. Whoever builds it, owns it. Whoever owns it, governs it. Europe’s infrastructural future must encode democracy, sustainability and human dignity.
The EuroStack embeds those values into 21st-century foundations. The infrastructure is already taking shape. The question isn’t whether it’s possible — it’s whether it happens on systems we control democratically or on infrastructure controlled by interests opposed to our own.
