90% of AI models and algorithms inside organizations are unknown to the organization itself. When infrastructure goes down in an agentic world, you don’t just lose connectivity — you lose context, while autonomous agents may already be mid-conversation with other systems. This panel brings together cybersecurity veterans, AI governance experts, and a Silicon Valley entrepreneur to debate digital infrastructure risk in 2026.─────────────────────────────WHAT THIS PANEL COVERS→ Why AI should be treated like an adolescent: powerful but not yet mature enough for unsupervised access — parenting, not relying→ The hidden SaaS surveillance problem: how embedded AI models inside enterprise software are reading audio files and selling functional data without corporate knowledge→ Why the European Commission’s €200 billion tech subsidy program flows to B2B companies when European values have always centered on the middle class→ The case for data sovereignty: why Will.i.am told a Stanford/MIT audience to stop giving tech companies their data, and why even the biggest extractors don’t know you well enough for precision action→ How hybrid and private cloud architectures are resurging as companies realize that diversification — not consolidation — is the path to resilience─────────────────────────────PANELISTS🎙 Mike Butcher — Technology Journalist & Founder, PathFounders (Moderator)🎙 Sebastian — Founder, AI and Partners🎙 Tammy Shuring — CEO & Co-Founder, Polymathic🎙 Additional panelists on AI infrastructure and cybersecurity─────────────────────────────unDavos is a community-driven summit running during WEF week in Davos, democratizing the conversation around global challenges.🌐 undavos.com─────────────────────────────Tags: digital infrastructure, AI governance, cybersecurity, cloud computing, data sovereignty, agentic AI, operational stability, data privacy, enterprise security, cloud concentration, hybrid cloud, SaaS, AI safety, tech regulation, European tech, data centers, unDavos, Davos 2026, WEF---TRANSCRIPTIt's all in the pronunciation, he's founder of AI, founder of AI and Partners. Sebastian is lost in Davos somewhere. And finally, we have Tammy Shuring, who's CEO and co-founder of Polymathic, right. So, right, yeah, okay, let's just get my notes up here, sorry everyone, it's supposed to be here. Digital infrastructure is now part of the operational background of industrial systems. We're going to be looking at cloud concentration risk, cyber security, you name it. Now, one of the issues, let's kick this off, when digital infrastructure fails today, what breaks first? Is it IT systems, production lines, safety, or is it all manner of things these days? Could be anything, couldn't it? What are the most common scenarios? I think for a big part, it's already broken. We don't know really what is in addition to infrastructure, so what we do is we identify AI systems basically within cloud, on-prem, and SaaS, and we do that per department. And what we really see is 90% of all the models and LAMs and algorithms that we identify are unknown by the organization. And if you don't know what it is, what works then? Right. I can be very short, I don't think there is a predefined order. Could you switch your microphone on? Switch it on. Now better? Yeah. Okay. So, I think it can happen simultaneously, right? Also, maybe the question is more not what breaks first, but what has the greatest impact, and that again depends on the industry. I mean, it's definitely potentially detrimental if that breaks. Yeah, my opinion is the biggest thing to really watch out for is you're going to lose all context, and in a world that's becoming progressively agentic, we're going to be live in the moment taking action on behalf of systems, groups, companies, people, and if infrastructure, when infrastructure goes down, we have no ability to have context, yet these agents could already be in conversations with other agents, other systems, and s