What Europe teaches the world about technology
Every piece of software has a personality, defined by its architecture: what data it collects and from whom, who gets to see it, what it optimizes for, and what it is willing to sacrifice for speed. An algorithm trained in Shenzhen carries different instincts than one trained in San Francisco, which behaves differently again from one built in Amsterdam. For most of the internet’s history, this was an observation you could make at dinner parties and then forget about. The underlying platforms were mostly American, the rules were mostly absent, and the rest of the world mostly went along with it.
That era is ending, in the space of about eighteen months. The United States, long the default custodian of the open internet, is dismantling the very institutions that made its tech ecosystem trustworthy. China continues to perfect its model of engineered convenience under state control. And Europe, perpetually mocked for regulating things it didn’t build, is suddenly the only major power still making a coherent argument for technology that serves democratic life.
Three civilizational models. Three sets of assumptions about what technology is for. And for the first time, all three are diverging fast enough that the differences will reshape not just apps and platforms, but alliances, infrastructure, and digital sovereignty.
The United States: When the referees join the team
In the summer of 2025, the U.S. General Services Administration added AI products from OpenAI, Google, and Anthropic to its government-wide purchasing schedule. OpenAI’s enterprise tools were pushed out to every federal agency that wanted them for a nominal fee of one dollar per year. The Department of Defense rolled out a commercial AI system for military planning. On the surface, this looked like modernization: government catching up with the private sector, a familiar and reasonable story.
Look closer and the picture changes. Around the same time, the administration signed executive orders directing federal agencies to review and potentially challenge state-level AI safety laws. States that passed their own regulations risked losing federal broadband funding. Agencies were instructed to procure AI only from developers of “ideologically neutral” models, a phrase with no technical definition but an unmistakable political one. The budget for the Cybersecurity and Infrastructure Security Agency, the body responsible for defending American digital systems, faced cuts of nearly $500 million and over a thousand positions. The State Department bureau that had spent years funding internet freedom tools for activists living under authoritarian regimes was gutted entirely: over 1,350 employees dismissed, more than 300 bureaus merged or eliminated, and flagship digital rights programs dissolved.
None of these moves, taken individually, would rewrite the story of American tech. Taken together, they describe a structural change. The “old” U.S. model, libertarian, messy, hands-off, had a kind of internal logic. Government stayed out of the way, companies moved fast, and if things broke (privacy scandals, algorithmic radicalization, foreign election interference), they were treated as bugs to be patched, not features to be designed around. It was reckless, but it was at least idealistic in its recklessness. What is replacing it is different. Instead of deregulation, it is selective regulation: protecting favored industries while suppressing oversight.
Rather than staying out of the way, the government and a handful of corporate giants fuse their interests so tightly that the boundary between public power and private platform is dissolving.
The country that built the open internet is now trending, by its own democracy indices, toward something it spent decades warning others about. And its tech policy is both symptom and accelerant. The U.S. will almost certainly remain the world’s dominant force in raw AI capability and platform scale for years to come. But its credibility as a neutral steward of digital infrastructure will continue to erode. The first signs are already here: the International Criminal Court in The Hague replacing Microsoft with a European open-source suite after a prosecutor was locked out of his email following U.S. sanctions. Denmark’s government piloting open-source replacements for American office software. Estonia accelerating an “open-source first” strategy driven by what its minister for digital affairs called “a matter of national survival”. These are obvious risk management efforts, and I expect they will multiply.
China: Pragmatism and the tech super-state
Somewhere in a ByteDance office several years ago, a team of thousands sat tagging short videos by hand, categorizing content frame by frame so that an algorithm could learn what holds human attention. The work was tedious, resource-intensive, and unglamorous. It was also one of the most consequential investments in consumer technology ever made. The recommendation engine that emerged from that effort powers TikTok, an app so effective at capturing attention that the Chinese government now classifies the underlying algorithm as a protected export, restricted by law from being sold or transferred abroad.
This is Chinese tech culture in miniature. A pragmatic focus on outcomes, backed by a state willing to pour resources into any project that serves national interests. WeChat started as a messaging app and became the operating system for Chinese daily life: payments, food delivery, government services, healthcare appointments, all inside one interface. The convenience is extraordinary. The trade-off (pervasive data collection, state access, no meaningful opt-out) is treated not as a scandal but as an acceptable cost of conveniency.
This model is well understood by now, and in its domestic form it hasn’t fundamentally changed. What did change is its bigger relevancy in a global conversation. As the United States undermines its own internet freedom credentials and attaches increasingly visible political strings to tech partnerships, the distance between Washington and Beijing on technology governance is narrowing. The two models are different. China’s surveillance apparatus, social credit experiments, and censorship infrastructure are unlike anything in the West. But the comfortable narrative of “authoritarian China versus free-world America” is getting harder to deliver with conviction. And into that narrative gap, China is quietly extending its tech governance model to the Global South by offering infrastructure, platforms, and financing with fewer apparent conditions than the U.S. now demands.
The real contest between America and China over the next decade will likely be about who sets the default terms for how technology is governed in the vast majority of the world that is neither American nor Chinese. And that contest is exactly where Europe’s role becomes critical, if Europe can deliver more than regulation.
Europe: The argument that needs an engine
Ninety percent. That is the share of Europe’s digital infrastructure (cloud, compute, core software) currently controlled by non-European companies, predominantly American ones. Amazon, Microsoft, and Google alone hold more than 70% of the European cloud market. This is the number that every hopeful story about European tech values has to contend with before it can be taken seriously.
Because Europe does have a compelling story. GDPR reshaped global privacy norms by forcing any company that touches European citizens’ data to comply with European rules. The AI Act is the first comprehensive regulatory framework for artificial intelligence anywhere on earth. The Digital Services Act and Digital Markets Act have begun to constrain platform monopolies in ways no American legislation has attempted. Products like Proton, Fairphone, and Mastodon are working proofs of concept for technology that respects privacy, sustainability, and community governance by design. The European Commission runs its own Mastodon instance to communicate with citizens. Privacy-focused search engines like Qwant and climate-conscious ones like Ecosia exist and function. None of this is trivial.
But writing rules for other people’s technology does not simply lead to sovereignty. Europe learned this the hard way with Gaia-X, a European cloud initiative that was supposed to reduce dependence on American hyperscalers. American companies lobbied to be included in the project. Once Microsoft, Google, and AWS were inside, the initiative lost its purpose. The continent produces more new startups than the U.S., but it cannot scale them into global players. The venture capital ecosystem remains fragmented. Brain drain is relentless. And saying there’s an AI spending “gap” is a gross understatement: U.S. hyperscalers are projected to spend over $700 billion on AI infrastructure in 2026 alone. Europe’s total sovereign cloud spending for the same year is forecast at roughly €10.6 billion. As the CEO of Mistral AI observed: the U.S. is building a new Apollo program every year, and you cannot regulate your way to computing supremacy.
So, what makes the current moment different from any previous round of hand-wringing? In November 2025, out of strategic necessity, all 27 EU member states signed a Declaration for European Digital Sovereignty. France and Germany convened a dedicated summit in Berlin. The word that keeps surfacing in official statements is one that would have seemed melodramatic five years ago: “survival.” Estonia’s minister for digital affairs called sovereignty “a matter of national survival, not just IT policy”. The EU’s 2026 legislative program is stacked accordingly: the Cloud and AI Development Act, the Quantum Act, the European Innovation Act, the Digital Fairness Act. Spending on sovereign cloud infrastructure is forecast to more than triple by 2027. Denmark is transitioning government offices to LibreOffice. Estonia is accelerating an open-source first strategy. The ICC has replaced Microsoft entirely with a European open-source suite built by a German state-owned organization.
The question is whether this urgency translates into execution, or whether it dissipates into another cycle of summits and declarations. The honest answer is that Europe will most probably not build a European AWS. But Europe does not need to replicate Silicon Valley’s model to achieve meaningful sovereignty. The genuine opportunity is architectural: building on digital public infrastructure, open standards, and interoperability rather than proprietary platforms. The European model is richer if it is built on SMEs rather than on American-style hyperscalers.
The most likely outcome over the next five years is messy but meaningful. Europe will reduce its most dangerous dependencies in government and critical infrastructure while remaining deeply reliant on U.S. hyperscalers for commercial workloads. But even partial sovereignty, if built on genuinely open architecture, would represent something dozens of other democracies are desperate for: a third model of technology governance, aligned with neither Washington nor Beijing, but interoperable with both.
What this actually means
Every technology carries the DNA of the society that built it. For most of the internet age, that society was American, and the DNA was optimistic, commercial, and allergic to rules. Now three distinct strands are diverging. One increasingly authoritarian despite its democratic branding. One authoritarian by design and getting better at exporting itself. And one democratic in principle but struggling to build the infrastructure its principles require.
Europe’s lesson for the world is not that regulation beats innovation. Europe innovates plenty. It simply optimizes for durability and trust over speed and scale, which is a different bet, not a lesser one. The real lesson is more uncomfortable for everyone involved. For Americans: the institutions that made your tech ecosystem trustworthy are being dismantled from within, and the rest of the world has noticed. For China: the walled garden works for control, but it does not export well to populations that have experienced democratic governance. For Europeans: the best argument in the room means nothing if you cannot build the engine to back it up.
The question for the next decade is which model of technology governance proves sustainable: economically, socially, and democratically. Europe has the most coherent answer. Whether it has the execution to match is the open question of the decade. And unlike previous rounds, this time there may not be a next round.
The infrastructure decisions being made right now, in data centers, legislative chambers and git repositories, are pivotal. Get them right, and Europe becomes the credible third path the world is looking for. Get them wrong, and sovereignty stays where it is now: in a PowerPoint deck.