This episode we roast three AI startups that recently closed rounds. Axiom Math wants to build a superhuman math reasoning system — great team, no product yet, and we explain why that's a problem. Nexus is an enterprise agent platform riding the AI wave, but in a market so saturated that distribution beats technology every time. And Ivo, the Australian contract intelligence startup that just raised a $55M Series B, is turning legal contracts into business intelligence — real traction, real risks, and a moat that could cut both ways.

This is a special episode on ICP — Ideal Customer Profile — the one thing every founder needs before anything else in lead-gen. We cover why "selling to everyone" is a fear response, not a strategy, and how to move from vague demographics to a psychographic profile that actually drives messaging. We walk through three diagnostic questions that reveal what your customer Googles at 3 AM, what they've already tried, and why it failed — then apply the framework live to a real startup: a YouTube transcriber tool built by the co-host. By the end, you've got a named persona, a clear pain stack, and messaging that writes itself.

In this episode, we roast two European startups and three investment funds — and end up noticing that most of what's being funded right now isn't startups at all. Upwest (Berlin) wants to be the investment operating system for legacy banks, but AI is collapsing the very complexity they're paid to abstract away. Kupando (Schönefeld) is betting on innate immunity therapy KUP101 for cases where standard immunotherapy fails — biotech plays by completely different investment rules, and that's worth understanding. We also briefly cover Partech Impact Fund, Montis VC, and Homaio: three ESG-flavored vehicles where the real story is about choosing sides, not backing innovation.

A special episode dedicated entirely to market research — how it worked before AI, why it was broken, and how LLMs have completely changed the game. We walk through the full six-step market analysis process side by side: traditional methods versus AI-powered tools, from hypothesis generation with Claude to synthetic discovery interviews with Atypica.ai and Quantilope, real-time competitive monitoring with Crayon and Brandwatch, and AI-generated report synthesis that removes analyst bias. The bottom line: a founder can now complete a full market analysis in a single day for about $100 in subscriptions — and it should no longer be a one-time project, but a continuous loop feeding every custdev cycle.

Europe's AI infrastructure race is heating up — N-Scale just raised €2 billion plus a €1.1B credit line, while Isembard pulled in £37.5M for AI-powered factory operating systems targeting aerospace and defense. We also dig into Lemrock, a French startup building commerce infrastructure inside ChatGPT and Claude via MCP servers — and share why LLM marketplaces are the next big distribution wave that most founders are still sleeping on. Rounding it out: Outpost's merchant-of-record platform and when it actually makes sense to use one.

This episode is about the gap between a compelling pitch and an actual product. We look at four funded startups: Tower (Berlin), building a last-mile AI platform for data engineers that still looks a lot like an agency; Tangled, a decentralized GitHub alternative that solves a problem most developers don't have; BrainCheck, a cognitive health testing tool that employers can use to catch burnout early — ethically complex but commercially sound; and Sensor Systems, solar-powered AI cameras for construction sites that are part of a real and fast-moving trend to automate the control layer on top of physical work.

This week's theme: why the best AI startups win through partnerships, not just vertical focus. We break down four companies — Escape (AI offensive security), Wayve (autonomous driving stack), Plato (sales automation for B2B distributors), and Biorce (AI for clinical trials). Sharp takes on business models, distribution strategy, and exit scenarios.