Adoption without foundation — Europe is deploying AI faster than it's settling the rules

The most consequential 24 hours in European AI industry news since the Act came into force. In Copenhagen, the WHO's first-ever EU-27 snapshot confirms 74% of Member States are already using AI diagnostics and 63% run patient chatbots. In Hannover, Chancellor Merz and Siemens CEO Busch deliver, in back-to-back appearances, the most direct industry challenge to the AI Act yet. In London, Boehringer Ingelheim opens a £150M AI centre — and a four-month-old startup raises $500M at a $4 billion valuation. Every one of these stories is real, verified, and dated yesterday or today. Seven days to the trilogue.

On Monday in Copenhagen, the WHO Regional Office for Europe published its first dedicated assessment of artificial intelligence adoption across the 27 EU Member States. Produced under a multi-year funding agreement with the European Commission, the report builds on WHO/Europe's broader regional survey published in late 2025 — but this time zooms in on the EU specifically. The data was collected between June 2024 and March 2025.

The headline numbers: 74% of EU countries are already deploying AI in clinical diagnostics — imaging, disease detection, clinical decision support. 63% run AI chatbots for patient engagement. 81% involve stakeholders in AI governance discussions. Nearly half have created dedicated professional roles for AI and data science inside their health ministries. Every one of the 27 identifies improved patient care as a primary driver.

Speaking to Bloomberg at Hannover Messe 2026 on Monday, Siemens CEO Roland Busch said most of the company's €1 billion industrial AI investment will be directed to the United States. His reasoning was blunt: the EU AI Act and Data Act, he argued, "miss the mark" by treating industrial and machine data with the same stringency as personal consumer data, layering new oversight on top of sector-specific regulations that already exist.

"It's complete nonsense to treat industrial and machine data the same way as personal data," Busch said. "I can't explain to my shareholders why I'm investing money in an environment where I'm being held back." The day before, Chancellor Friedrich Merz used his opening speech to back Busch's position: the German government, he said, will push "for extricating industrial AI from the current, overly restrictive straitjacket of the EU's regulatory framework."

Boehringer Ingelheim, the 141-year-old German biopharma company, officially opened its fourth Computational Innovation centre on Monday — this one at King's Cross in London's Knowledge Quarter, alongside Google, OpenAI, and Meta. The company has committed £150 million over ten years, with the first 50 AI experts expected in place by the end of 2027. UK Science Minister Lord Patrick Vallance spoke at the launch event.

The choice of location is the story. Boehringer already has Computational Innovation hubs in Austria, Germany, and the United States. The fourth was always going to be politically meaningful. "The UK has a strong legacy in AI, and the government's continued commitment to advancing data-driven innovation in life sciences and healthcare makes it an ideal location," said Paola Casarosa, Boehringer's Global Head of Innovation Unit. The investment also lands just months after US pharma giant MSD scrapped its own £1B UK expansion in September 2025 — Boehringer is, in effect, filling exactly the gap an American company left.

Recursive Superintelligence was incorporated in London on 31 December 2025. Four months later, the company has raised at least $500 million at a $4 billion pre-money valuation, led by GV (formerly Google Ventures) with participation from Nvidia. The round was so oversubscribed it could close at $1 billion. The founding team: Richard Socher (ex-chief scientist, Salesforce), Tim Rocktäschel (AI professor at UCL, ex-principal scientist at Google DeepMind), and former OpenAI researchers Josh Tobin, Jeff Clune, and Tim Shi. Twenty staff. No product. A public launch is targeted for mid-May.

The ambition is "recursive self-improvement" — AI systems that can continuously improve themselves without human intervention. The choice of London incorporation is what makes this a European story. Asked by Implicator.ai why he did not build the company in continental Europe, Socher did not mince words: the AI Act, he said, has "slowed the whole region down even more than it already was," stacking bureaucracy on bureaucracy. His line on what founders now say about Europe: "the most beautiful open-air museum."

Last week's spotlight fell on Cyprus — the country that will broker the trilogue. This week's falls on Germany — the country making the argument against the Commission's own proposal. In fewer than 48 hours at Hannover Messe, Chancellor Friedrich Merz and the CEOs of Germany's three most valuable tech-adjacent companies — Siemens, SAP, and Deutsche Telekom — lined up on the same stage to press the same point: the AI Act, as drafted, is pushing European industrial AI investment overseas.

The Merz position is now formal German government policy. Compute capacity doubling by 2030. Industrial AI carve-out from Article 50 transparency. Separate treatment for machine and industrial data. Merz arrives at the Nicosia informal European Council on Thursday with this stack — just five days before the Strasbourg trilogue.

What to watch: whether Merz signals a German veto threat on the Omnibus if industrial AI is not meaningfully exempted. He has not done so publicly. He does not need to. Hannover did it for him.