A thought experiment published by Brussels-based think tank Arq Foundation imagines a dark future where Europe has been economically gutted by American and Chinese AI dominance. Titled 'Europe 2031,' the scenario went viral among policymakers after being read by members of the European Parliament and discussed in track 1.5 talks between British and German officials. The timing proved uncomfortably prophetic: it dropped one day before the Trump administration moved to block "foreign nationals" from accessing Anthropic's Claude Fable model, validating at least one of its core predictions about US restrictions on advanced AI.

The Doomsday Blueprint

The scenario unfolds through fictional Brussels staffer Caroline Dubois and her San Francisco-based German friend Christian Vogt. What starts as a buddy comedy about cultural differences—Dubois is baffled by America's "70 or 80-hour" working weeks and tech-bro conviction that everything is about to change—devolves into geopolitical catastrophe. By 2031, the US has monopolized roughly 70% of global compute capacity through massive datacentre investments, while Europe dithers on AI adoption and ignores advisors' pleas for regulatory carte blanche for datacentre providers. The EU's one bargaining chip becomes ASML, the Dutch lithography firm essential to semiconductor production—but even that proves insufficient leverage against frontier AI spyware deployed by Washington.

Real Deals Already Crumbling

Here's where it gets interesting from a hacker perspective: several billion-dollar deals the authors cite as evidence of America's AI push have already collapsed. The $100bn OpenAI-Nvidia agreement, dubbed "the biggest AI deal of last year," evaporated in February. The $300bn OpenAI-Oracle partnership looks increasingly doubtful as reports indicate OpenAI remains billions underwater while burning cash on infrastructure. Even the bulldozers breaking ground in Texas for OpenAI's flagship datacentre project have reportedly stopped turning. Yet co-author Maximilian Negele, a former Rand analyst who left his job this year to focus on the project, argues these are mere bumps in the road. "I wouldn't rule out that there's some exuberance and that one or two AI companies might go bankrupt," he told The Guardian. "But what we wanted to get across is a general feel for a version of what we think will happen."

The Translation Problem

Negele says he wrote Europe 2031 to bridge the "incredible translation barrier" between Brussels and San Francisco, where AI is being developed at breakneck speed. His co-author Alex Petropolous frames it as a supply-side problem: only so many datacentres can be built globally each year, and Europe needs to claim its share before it's too late. The authors advocate for "AI zones" where power and planning regulations would be streamlined and deregulated—a tough sell in an era when, as Petropolous notes, "people hate AI in general... people hate datacentres. They destroy the landscape." MEP Nicolás Casares from Spain offers a more skeptical read: while he acknowledges some scenarios could unfold, he thinks the authors are "increasing—[just] a bit—the alarms in order to call our attention."

What Europe Should Actually Do

Casares cuts through the alarmism with sharper questions: what value does OpenAI or Anthropic infrastructure bring to Europe if American policymakers can simply flip the switch? "We are buying a narrative that we need a lot of datacentres not to lose the race for AI," he said. "But this is crazy... we are paving the way for infrastructure that they will use and sometimes not allow us the possibility of using it." The scenario's authors, notably coy about their funding sources—Arq Foundation describes itself as "neither an advocacy NGO nor a venture-backed startup" but discloses nothing—have nonetheless crystallized a conversation Europe desperately needed to have.

Key Takeaways

  • Europe 2031 validates fears that US AI restrictions could target European users, as demonstrated by the Fable blockade
  • Multiple billion-dollar US AI infrastructure deals cited in the scenario have already collapsed or stalled
  • The core debate isn't whether AI will transform economies—it's who controls that transformation and on whose terms
  • MEPs are now questioning whether building American-owned datacentres actually constitutes EU tech sovereignty