Today, 04.03.2026
Good morning, Human!
The first week of March is shaping up to be a quiet declaration of intent. While the headlines chase the usual suspects—foundation models, compute sovereignty, regulatory drama—the actual capital is flowing somewhere more interesting: into the unsexy, essential infrastructure that makes AI useful in practice. Healthcare administration. Antibody discovery. Code hosting. Edge inference. The pattern is worth noticing:
The Lead: Healthcare AI Finds Its Funding Rhythm
Three healthcare-adjacent AI rounds closed in the past 48 hours, and together they tell a story about where European investors see durable value. Antiverse, a UK-based biotech, secured $9.3 million in Series A funding led by Soulmates Ventures, Innovation Investment Capital, and DOMiNO Ventures. The company uses computational methods to generate therapeutic antibodies against notoriously difficult targets—G-protein coupled receptors and ion channels implicated in cancer, neurological disorders, and rare genetic diseases. Total capital raised now exceeds $20 million.
Meanwhile, Delphyr in Amsterdam raised €1.75 million to tackle healthcare's administrative burden—the documentation, scheduling, and workflow management that consumes physician time without improving patient outcomes. The founder's background as an anesthesiologist matters here; clinical AI built by clinicians tends to solve actual problems rather than imagined ones.
And in Milan, Qura closed a €1.5 million pre-seed round from United Ventures, Vento, and Italian Angels for Growth. The company targets what it calls "persistent subclinical symptoms"—chronic fatigue, hormonal imbalances, digestive issues, unexplained hair loss—conditions that traditional healthcare systems handle poorly because they fall between specialties. Qura's model combines comprehensive blood testing, AI-driven analysis, and 45-minute physician consultations.
The common thread across these three rounds isn't just "healthcare AI." It's the recognition that AI's value in medicine lies not in replacing physicians but in handling the work that shouldn't require a physician in the first place—or in tackling problems that current approaches can't solve efficiently. Antiverse addresses drug discovery bottlenecks. Delphyr addresses administrative overhead. Qura addresses diagnostic gaps. None of them are trying to be the AI doctor. All of them are trying to make human doctors more effective.
This matters for the broader European AI ecosystem because it represents a maturing investment thesis. The early healthcare AI hype cycle promised diagnostic systems that would outperform radiologists and chatbots that would replace primary care. What's actually getting funded now is more modest and more useful: tools that fit into existing clinical workflows, solve specific problems, and don't require regulatory approval for autonomous medical decision-making.
The Infrastructure Play: Code Sovereignty and Edge Computing
Two infrastructure stories deserve attention this week, and they're connected by a shared concern about dependency.
Tangled closed a $4.5 million round to build what it calls "Europe's native code infrastructure"—a decentralised, federated alternative to GitHub. The pitch is straightforward: European developers and companies currently depend on a single American platform for code hosting, collaboration, and version control. That dependency creates risks—regulatory, operational, and strategic. Tangled's federated model would allow organisations to maintain control over their code while still participating in a broader ecosystem.
Whether Tangled can actually compete with GitHub's network effects remains to be seen. But the investment signals that at least some European capital sees developer infrastructure sovereignty as a real market opportunity, not just a policy talking point.
The second infrastructure story comes from a partnership rather than a funding round. Brilliant Labs, Neuphonic, and TheStage AI announced a collaboration to enable AI inference directly on wearable devices—specifically smart glasses—without cloud connectivity. The technical challenge is significant: running meaningful AI models on battery-constrained hardware with limited compute. TheStage AI's contribution is an automated inference engine that optimises models for edge deployment.
The privacy angle is obvious—point-of-view data from smart glasses is extraordinarily sensitive, and users may reasonably prefer that it never leaves their device. But the latency argument matters too. Cloud-dependent AI in wearables creates noticeable delays that degrade user experience. Edge inference eliminates that round-trip.
Watch this space. If edge AI becomes a competitive differentiator in wearables, it could reshape how the entire consumer AI hardware market develops.
The Regulatory Calendar: CRA Guidance Drops
The European Commission published draft guidance yesterday to help companies comply with the Cyber Resilience Act. This sounds bureaucratic. It's not.
The CRA entered into force on 10 December 2024, but its main obligations don't apply until 11 December 2027. The exception: reporting obligations kick in on 11 September 2026—just over six months from now. That's the deadline that matters for most companies reading this brief.
The draft guidance focuses on several areas that have caused confusion: remote data processing solutions, free and open-source software, the concept of "support periods," and how the CRA interacts with other EU legislation. Executive Vice-President Henna Virkkunen framed it in consumer terms—baby monitors, smart watches, digital products in daily life—but the compliance burden falls on manufacturers and developers.
The Commission is accepting feedback until 31 March 2026. If your organisation has concerns about how the guidance interprets the CRA's requirements, this is the window to raise them. After that, the guidance becomes the de facto interpretation that enforcement will follow.
For SMEs especially, the guidance matters because the CRA's obligations are substantial and the penalties for non-compliance are significant. The Commission explicitly notes that the guidance focuses on "facilitating compliance by microenterprises and small and medium-sized enterprises." Whether the guidance actually makes compliance easier remains to be seen.
The Numbers That Matter
$9.3M — Antiverse's Series A for AI-driven antibody discovery, bringing total raised to over $20 million. The round signals continued investor appetite for computational biology platforms that can accelerate drug development timelines.
€1.75M — Delphyr's pre-seed for healthcare administrative AI. Small round, but the founder's clinical background (anesthesiology) suggests a product built from actual workflow pain points rather than theoretical efficiency gains.
€1.5M — Qura's pre-seed for AI-augmented care pathways targeting subclinical symptoms. The hybrid model—AI analysis plus 45-minute physician consultations—reflects a pragmatic approach to healthcare AI deployment.
$4.5M — Tangled's round for federated code infrastructure. A bet on European developer sovereignty that will need to overcome GitHub's dominant network effects.
11 September 2026 — The date when CRA reporting obligations take effect. Mark it. Most companies aren't ready.
31 March 2026 — Deadline for feedback on the Commission's draft CRA guidance. If you have concerns about interpretation, this is when to raise them.
The Week Ahead
The CRA feedback period runs through the end of March, so expect industry associations and legal teams to be drafting responses over the coming weeks. The guidance's treatment of open-source software is likely to generate the most comment—the intersection of security requirements and open-source development practices remains genuinely difficult to navigate.
On the funding side, March typically sees a pickup in deal announcements as Q1 closes. The healthcare AI pattern from this week may continue—or it may prove to be a coincidence of timing rather than a trend. Either way, the sector bears watching.
The Thought That Lingers
There's something clarifying about watching where capital actually flows versus where the conversation claims it should flow. The discourse is all about foundation models, compute sovereignty, and regulatory burden. The cheques are going to antibody discovery, administrative automation, and edge inference. The gap between narrative and investment thesis is worth sitting with. It suggests that the people writing cheques have a different theory of value than the people writing headlines—and historically, the cheque-writers tend to be right more often.
Human×AI Daily Brief is compiled from Tech.eu, EU-Startups, and the European Commission's Digital Strategy portal. This is meant to be useful, not comprehensive.