Today, 28.04.2026
Good morning, Human. The week opens with Brussels telling Google exactly how its Android operating system must open up to rival AI services, while across the Channel, a former DeepMind researcher just raised more money at seed stage than most European startups see in their entire lifecycle. These two stories might seem unrelated. They are not. Both reveal the same underlying tension: who gets to control the infrastructure layer of the AI era, and on what terms.
In Brief
The European Commission has published draft measures requiring Google to let competing AI services access Android's core capabilities on equal terms with Gemini, including custom wake words and deep app integration. This matters because it tests whether the Digital Markets Act can actually shape AI competition before market positions calcify. For Europe, the consultation deadline of 13 May 2026 offers a narrow window for AI developers to influence rules that could determine whether European AI services can compete on the dominant mobile platform.
This brief scratches the surface. The real conversation about Europe's AI future happens in person, May 19 in Vienna, at Human x AI Europe, where the people building this ecosystem gather to figure out what comes next.
The Lead: Brussels Tells Google How to Open Android for AI
The European Commission yesterday published its preliminary findings in the specification proceedings against Google under the Digital Markets Act, and the document reads like a technical manual for dismantling platform advantage. The Commission wants Google to allow competing AI services to be triggered using custom wake words, a capability currently reserved exclusively for Gemini. It wants third-party AI assistants to access the same deep integration with apps that Google's own services enjoy: sending emails, ordering food, sharing photos, adjusting system settings.
The scope is striking. The draft measures cover four capability areas: invocation features (how users activate AI services), context features (how AI services access on-device data), action features (how AI services interact with apps and system settings), and resource features (access to on-device AI models and processing power). Google would be required to provide these capabilities free of charge across all Android devices, including those from third-party manufacturers, through documented APIs and frameworks.
Executive Vice-President Teresa Ribera framed the stakes clearly: AI services are becoming more and more relevant for EU citizens' daily interaction with their mobile devices. And it is critical to protect innovation by AI companies of all sizes. Her colleague Henna Virkkunen added that interoperability is key to unlocking the full potential of these technologies.
Google's response was predictable but pointed. The company called the intervention unwarranted and warned it would mandate access to sensitive hardware and device permissions, unnecessarily driving up costs while undermining critical privacy and security protections for European users. The privacy argument is not frivolous: giving third-party AI services deep access to device data and app interactions does create new attack surfaces. The question is whether that risk outweighs the competition benefits.
The consultation runs until 13 May 2026, with a final decision expected by 27 July. For European AI developers, this is a rare opportunity to shape the rules of engagement on the platform that dominates mobile computing. For Google, it represents the DMA's most direct intervention yet into how AI services compete at the operating system level. The outcome will signal whether Europe's competition framework can move fast enough to matter in a market where positions are hardening by the quarter.
The Funding Picture: Ineffable's $1.1 Billion Seed Rewrites the Playbook
David Silver, the former head of reinforcement learning at Google DeepMind, emerged from stealth yesterday with Ineffable Intelligence and a $1.1 billion seed round at a $5.1 billion valuation. The numbers are so large they risk obscuring what they actually represent: Europe's largest seed financing ever, by a considerable margin, for a company pursuing superintelligence through reinforcement learning rather than the large language model paradigm that dominates current AI development.
The investor list reads like a who's who of AI capital: Sequoia and Lightspeed co-led, with participation from Nvidia, Google, DST Global, Index, and the UK's Sovereign AI Fund. The British Business Bank invested $20 million directly, adding Ineffable to a portfolio that includes Wayve and PolyAI. UK Science and Technology Secretary Liz Kendall called it evidence of our determination to ensure that the UK isn't just an AI taker but an AI maker.
Silver's thesis is a direct challenge to the prevailing approach. Where most frontier AI labs train models on internet text, Ineffable aims to build what it calls a superlearner that discovers knowledge through its own experience, from basic motor skills to intellectual breakthroughs. This is the paradigm that produced AlphaGo, AlphaZero, and AlphaFold, all projects Silver led at DeepMind. The bet is that reinforcement learning, not imitation of human-generated data, is the path to genuine artificial intelligence.
The timing matters. This round lands amid a broader exodus of top researchers from Big Tech labs to launch their own ventures. Recursive Superintelligence, founded by another former DeepMind engineer, is reportedly raising up to $1 billion. AMI Labs announced a $1 billion raise in March after Yann LeCun left Meta. The talent that built the current generation of AI systems is now building the next one, and they're doing it with unprecedented capital backing.
For Europe, the question is whether Ineffable's London headquarters represents a genuine shift in where frontier AI gets built, or whether the capital and talent will eventually migrate to where the compute and regulatory environment are more accommodating. The UK's Sovereign AI Fund participation suggests the government understands the stakes.
The Sovereignty Situation: Europe's Messy Breakup with US Tech
France's decision to migrate 2.5 million government computers from Windows to Linux, announced earlier this month, continues to ripple through European policy circles. The move is part of a broader directive from DINUM, France's digital agency, requiring every ministry to submit plans for eliminating extra-European digital dependencies by autumn 2026. The scope extends beyond operating systems to collaboration tools, antivirus software, AI platforms, databases, and network equipment.
Minister David Amiel's framing was unambiguous: The State can no longer simply acknowledge its dependence; it must break free. We must become less reliant on American tools and regain control of our digital destiny. The language of sovereignty has shifted from aspiration to operational mandate.
The European Commission's own cloud procurement offers a template. In April, the Commission awarded contracts to four providers under its new Cloud Sovereignty Framework: a Luxembourgish-French partnership led by Post Telecom with OVHCloud and CleverCloud, Germany's STACKIT (Schwarz Group), France's Scaleway (Iliad Group), and a Belgian-French-Luxembourgish partnership led by Proximus using S3NS (a Thales-Google Cloud joint venture). The framework introduces Sovereignty Effectiveness Assurance Levels (SEAL) ranging from SEAL-0 (no sovereignty) to SEAL-4 (full EU supply chain from chips to software). Most awarded providers reached SEAL-3, meaning their services cannot be blocked by non-EU third parties.
The EuroStack initiative, which advocates for a complete European digital stack, has gathered over 400 signatories from across the tech ecosystem. The European Parliament's January resolution directing the Commission to map and reduce foreign technology dependencies passed with 77% support. The political will exists. The question is execution.
Critics point to the gap between rhetoric and reality. Mark Scott, senior resident fellow at the Atlantic Council, called France's Windows migration more smoke and mirrors than anything else, noting that procurement contracts and IT dependencies take years to unwind. The French Court of Auditors has questioned spending on in-house tools like Visio, the government's replacement for Zoom and Teams. And private companies haven't followed the government's lead: Lufthansa and Air France both chose Starlink for in-flight Wi-Fi, despite the sovereignty concerns.
The deeper tension is structural. Even as governments replace Windows with Linux and Teams with Visio, the cloud and compute substrate beneath European AI development remains predominantly American. Replacing the desktop layer matters, but it sits above infrastructure that the sovereignty movement has barely begun to address.
The Numbers That Matter
$1.1 billion: Ineffable Intelligence's seed round, Europe's largest ever, at a $5.1 billion valuation. The previous European seed record was substantially smaller.
13 May 2026: Deadline for feedback on the Commission's proposed Android interoperability measures. Final decision expected by 27 July.
2.5 million: French government computers targeted for migration from Windows to Linux under DINUM's digital sovereignty directive.
70%: Share of European cloud market controlled by Amazon, Microsoft, and Google, according to industry estimates. Local providers hold approximately 15%.
SEAL-3: Sovereignty level achieved by most providers in the Commission's cloud procurement, meaning services cannot be blocked by non-EU third parties.
€180 million: Value of the Commission's sovereign cloud tender, awarded to four provider consortia in April.
The Week Ahead
1 May: Deadline for feedback on the Commission's separate consultation regarding Google Search data sharing with third-party search engines and AI chatbots.
13 May: Deadline for feedback on Android interoperability measures under DMA specification proceedings.
Late May: Expected publication of updated Cloud Sovereignty Framework based on lessons from the Commission's procurement tender.
June 2026: First Industrial Digital Meetings scheduled by DINUM to formalize public-private coalitions supporting France's digital sovereignty transition.
The Thought That Lingers
There's something almost paradoxical about the week's two biggest stories. In one, Europe's regulators are telling Google exactly how to open its platform to competitors, down to the technical specifications of wake words and API access. In the other, a British startup just raised more money at seed stage than most European AI companies will ever see, backed by the very American tech giants that European policy is designed to constrain. The investors in Ineffable include Google and Nvidia. The UK's Sovereign AI Fund sits alongside Sequoia and Lightspeed.
This isn't hypocrisy. It's the reality of building frontier technology in 2026. Sovereignty doesn't mean isolation. It means having options, having leverage, having the capacity to say no when the terms aren't acceptable. The Commission's Android intervention and France's Linux migration are about creating those options. Ineffable's funding is about proving that world-class AI research can happen in Europe, even if the capital comes from everywhere.
The question isn't whether Europe can decouple from American tech. It can't, not fully, not for a generation. The question is whether it can build enough independent capacity to negotiate from strength rather than dependence. That's what this week's news is really about.
Frequently Asked Questions
What exactly does the EU want Google to change about Android?
The European Commission wants Google to allow competing AI services to use custom wake words, access on-device data, interact with apps and system settings, and use on-device AI models - capabilities currently reserved for Google's own Gemini service. These changes would apply to all Android devices through documented APIs and frameworks, provided free of charge.
Why is Ineffable Intelligence's funding round significant?
At $1.1 billion, it's Europe's largest seed round ever, representing a major bet on reinforcement learning over large language models. Founded by former DeepMind head David Silver, it signals that world-class AI research can attract massive funding in Europe, even from American investors like Google and Nvidia.
What is digital sovereignty and why does it matter?
Digital sovereignty refers to a country's ability to control its digital infrastructure and reduce dependence on foreign technology providers. For Europe, it means building alternatives to American tech platforms and cloud services to maintain strategic autonomy and negotiating power in the digital economy.
How realistic is France's plan to replace Windows with Linux?
Critics call it ambitious but challenging, noting that migrating 2.5 million government computers involves complex procurement contracts and IT dependencies that take years to unwind. The broader question is whether replacing desktop software addresses the deeper infrastructure dependencies on American cloud and compute services.
Human×AI Daily Brief is compiled from European Commission announcements, CNBC, Reuters, TechCrunch, EU-Startups, Yahoo Finance, and primary regulatory documents. This is meant to be useful, not comprehensive.