- Superintelligence
- Posts
- Superintelligence Europe — No. 003
Superintelligence Europe — No. 003
Zero OpenAI trials in the UK. Europe's strategic ambiguity laid bare. Four AI funding rounds point to one theme. Your Monday briefing.
In partnership with
Everything that moved in European AI over the weekend · Across all 27 EU member states | ||||||||||||
| ||||||||||||
Issue No. 003 — Monday, 23 March 2026 The weekend brought a quieter news cycle — but not a quiet one. A Freedom of Information request in the UK exposed a yawning gap between AI partnership announcements and actual deployment. In Brussels, EU AI Week concluded after seven days of workshops on sovereignty, governance, and public-sector AI. Analysts sharpened the EU-vs-US regulatory contrast following Washington’s new light-touch AI framework. And a batch of European AI startups — from Zurich to Bucharest — closed funding rounds that point unmistakably toward one theme: AI agents moving into the physical world. This is European AI at the weekend: slower in volume, sharper in signal. The institutions are writing rules. The capital is moving. The gap between announced ambitions and operational reality is widening. Six stories. Your Monday morning briefing starts here. | ||||||||||||
| ||||||||||||
Lead · United Kingdom · Policy / Accountability 01The UK government ran zero trials of OpenAI technology — eight months after signing a high-profile AI partnership Source: The Guardian · 21 March 2026 A Freedom of Information request confirmed that the UK’s Department for Science, Innovation & Technology had not run a single trial or test of OpenAI models under the memorandum of understanding signed to deploy advanced AI across public services and tackle societal challenges. Eight months after that MoU was announced with considerable fanfare, the operational record is blank. Officials cited separate ChatGPT access via the Ministry of Justice and ongoing collaboration with the UK AI Safety Institute as evidence of broader engagement. But critics were unconvinced. Tarek Nseir of Vallance and Matt Davies of the Ada Lovelace Institute flagged execution gaps, transparency failures, accountability deficits, and the structural risk of long-term lock-in to US providers. A public opinion survey sharpened the political stakes: 84 percent of respondents said they were concerned that the government was prioritising industry interests over the public good. The Pattern That Worries Observers “The concern is not just about one MoU — it is about a pattern of announcements without accountability.” — Policy analysts cited in The Guardian’s coverage of the FOI findings Why It Matters for Europe Post-Brexit Britain is supposed to demonstrate that agile AI governance outside the EU framework can deliver faster results. Stories like this complicate that argument significantly. For Brussels, it is both a cautionary tale and, quietly, a validation: the EU’s slower, more procedural approach to AI deployment may frustrate innovators, but it makes this kind of accountability gap harder to hide. | ||||||||||||
Strategy · EU-wide · Geopolitics 02Trump’s AI playbook exposes Europe’s strategic ambiguity — and the cost of standing still Source: Modern Diplomacy · 21 March 2026 One day after Washington published its new AI framework — light-touch, innovation-first, minimal regulatory friction — analysts on Saturday crystallised the EU-US contrast in sharp terms. Europe’s risk-based, sovereignty-anchored approach built around the AI Act and GDPR was already well defined. What the Trump framework did was make the divergence impossible to ignore. The analysis circulating in policy circles over the weekend warned of compounding structural risks: European cloud and data infrastructure remaining materially dependent on US firms; EU companies carrying heavier compliance costs than their American competitors; and the risk that European regulatory standards influence global norms on paper while US deployment defines them in practice. The counterargument — that European regulation still anchors global debates on rights and accountability — was treated as real but fragile without stronger enforcement and domestic AI investment. The Structural Question Europe has rules. The question in 2026 is whether those rules translate into industrial leadership or merely compliance overhead for companies ultimately trained on US infrastructure, using US chips, and deploying US models. Strategic ambiguity on that question has a cost — and it compounds with time. | ||||||||||||
Legislation & Governance · EU-wide 03 Parliament backs AI Omnibus — with a new ban on AI nudification systems EU Parliament · PubAffairs Bruxelles · Week of 16–22 Mar The IMCO and LIBE committees adopted their positions on the digital omnibus — the package that would delay application of certain high-risk AI system rules and simplify provisions within the Act. The headline amendment that attracted most attention: an explicit ban on AI “nudification” systems, tools capable of removing clothing from images or generating non-consensual synthetic nudity. The position now advances toward trilogue negotiations with the Council. Parliament is simultaneously softening compliance timelines for businesses and hardening the prohibited-use list — a combination that reflects the political pressures pulling in both directions. 04 EU AI Week 2026 concludes in Brussels — sovereignty, agents & public-sector AI Belgium / EU · BOSA Belgium · 21–22 Mar The week-long series of open events on “Shaping EU AI Sovereignty” wrapped on Sunday. Belgian federal services (FPS BOSA) and AI4Belgium co-organised sessions spanning trustworthy AI, federated learning, cybersecurity, health applications, and AI for public administration. The closing sovereignty roundtable drew over 70 participants. The defining message of the final days: sovereignty is not merely a regulatory question but a capability and industrial one. Having rules is not sufficient. Having the compute, talent, and institutions to enforce and apply them is what determines whether Europe shapes AI — or is shaped by it. | ||||||||||||
Capital · Switzerland / Netherlands / Belgium / Romania · Funding Recap 05Europe’s AI agents move into the physical world — four rounds, four sectors, one clear theme Source: The Next Web weekly recap · 22 March 2026 The Next Web’s Sunday funding recap for the week of 16–22 March landed with a clear signal: European AI investment is converging on agentic systems operating in real-world, regulated environments — clinical trials, multilingual speech, blue-collar recruitment, and information integrity. These are the domains where earlier waves of automation largely failed to penetrate. The capital now flowing there suggests investors believe the moment has changed.
The Trend Beneath the Numbers All four rounds share a characteristic: they target sectors — pharma, multilingual services, blue-collar work, information integrity — where earlier automation waves failed to deliver. Investors appear to be betting that the agentic AI generation succeeds where previous software layers did not. GDPR compliance and European data sovereignty are appearing as product features, not regulatory burdens. That shift in framing matters. | ||||||||||||
Sovereignty · EU-wide · Defence / Infrastructure 06The EU plans a sovereign defence AI data space by 2030 — free from US infrastructure Source: Militarnyi / Euractiv · 22 March 2026 A European Defence Agency presentation seen by Euractiv outlined plans for the European Defence Artificial Intelligence Data Space (DAIDS) — a secure, interoperable data-sharing environment to support AI-enabled defence operations across member states, built explicitly without US infrastructure involvement. Contracts are already signed: a consortium of CEA (France), Cloud Data Engine, and Sopra Steria will build the platform, with integration into armed forces and defence authority operations planned for 2029–2030. The project sits within the EU’s broader Defence Roadmap to 2030. It represents one of the most concrete expressions yet of Europe’s push to build sovereign AI infrastructure in strategically sensitive domains — where dependence on external data environments is treated as a security risk, not merely a commercial inconvenience. What This Signals The DAIDS announcement is the clearest operational signal yet that Europe’s sovereignty rhetoric is moving into defence infrastructure. Civilian AI sovereignty is contested and slow. Defence AI sovereignty, where the strategic stakes are harder to dispute, may prove a faster forcing function. Watch whether commercial AI infrastructure follows the same logic. | ||||||||||||
07 · Signal — Verified Voices Credible accounts and publications that drove the weekend’s European AI conversation. Filtered for genuine signal. @thenextweb European tech media Published the definitive weekend funding recap for European AI. The most-amplified data-driven signal on where capital is moving in European AI this week. Remains the clearest aggregator of startup funding across the continent. @BelgiumBOSA Belgian Federal Digital Services Active throughout the final days of EU AI Week, sharing session updates, roundtable findings, and sovereignty framing. The most consistent official EU-level voice on AI governance events this weekend. Official posts are among the earliest signals of what Brussels is prioritising in practice. @AdaLovelaceInst UK AI research institute Cited in the Guardian’s OpenAI MoU story, the Ada Lovelace Institute’s commentary on the FOI findings drove significant weekend discussion on AI accountability and the structural risks of government-industry partnerships without transparent evaluation frameworks. |
If you work in fintech or finance, you already have too many tabs open and not enough time.
Fintech Takes is the free newsletter senior leaders actually read. Each week, I break down the trends, deals, and regulatory moves shaping the industry — and explain why they matter — in plain English.
No filler, no PR spin, and no “insights” you already saw on LinkedIn eight times this week. Just clear analysis and the occasional bad joke to make it go down easier.
Get context you can actually use. Subscribe free and see what’s coming before everyone else.