Venture Intelligence Day Prague, Report

If last year was “let’s test drive the Ferrari,” this year was “get in, we’re going for a ride.” Venture Intelligence Day returned to Prague with a clear theme: the edge in VC isn’t a single model; it’s data discipline + adoption. Panels ranged from risk to sourcing to platform ops, and the through-line was simple, buy what’s commodity; own what compounds. 

Our Sr. Partner & CTO, Sagar Chandna, attended the event to debate, discuss and exchange data-driven methodologies and ideas with of the top VCs in Europe.

The EIC Fund laid out how a public co-investor gets sharper with scale. They still require a qualified lead and won’t “make the market,” but with hundreds of deeptech investments on file they’re now debating terms from a stronger, data-driven footing. More interesting: they centralize cap-tables and run a “shareholder watcher” to flag growth in non-EU ownership - useful when your mandate includes European strategic autonomy. Earlybird reminded everyone that great risk teams find losers early, so the fund doesn’t accidentally average down. EnBW New Ventures reinforced that their incentives are financial-return first with a secondary strategic score, timing flexibility matters. Energy Revolution Ventures pushed the crowd to explicitly model commodity price shocks and policy/geopolitics because those swing hard tech theses overnight.  

What we’re taking back to RunwayFBU: institutionalize risk as data plumbing, cap-table registry + alerts, sector benchmarks for milestones/valuations, and quarterly follow-on reviews that are comfortable saying “no.”  

On deal sourcing, Balderton was refreshingly blunt: stop trying to train your own LLMs, off-the-shelf models plus strong prompt/workflow engineering are delivering the ROI (think: structuring decks and board packs, not AI-for-AI’ssake). NGP Capital still leans on traditional ML for tabular discovery, while using Gen AI downstream in portfolio ops. Speedinvest showed that a targeted scraper can sustain >10% of annual dealflow out of ~10k tracked leads - especially when you index on people data to spot stealth founders. And Dawn Capital demoed a two-year relationship knowledge graph (yes, it understands board seats and WhatsApp ties) that makes access more meritocratic and, frankly, wins deals. The adoption lesson? Deliver insights in the tools partners already open - Monday morning email, Slack, CRM. If it isn’t 10× easier, it doesn’t exist.  

RunwayFBU move: build a relationship graph + content pipeline; use LLMs for extraction, not magic; push a “Weekly 10” email of net-new founders with warm paths.  

The “smart money” bar moved, value before the term sheet now decides competitive rounds. The platform answer is to operate like a product team: one accountable owner per company, stage-based playbooks (Seed: founder productivity/hiring; A: PMF/design partners; B: leadership cadence), and a shared “asks” pipeline with SLAs so nothing falls through cracks. And if you’re a CVC, the strategic value is explicit: targeted research, white papers, BD intros, and co-investment pathways, alongside financial returns. Measure it with Founder NPS and referral-based dealflow.  

Giant Ventures gave a year one blueprint that actually ships: Reuse → Buy → Build (buy the CRM/portfolio app; build ELT + warehouse; layer thin “agents”). They run “Coffee” dashboards for deployment/company views and a warehouse powered chatbot-“Brian”-that drafts LP report first passes, acts as a dealflow historian (“have we seen this?”), and extracts/ routes portfolio asks so nothing gets dropped. None of this replaces people; it makes them faster and more consistent.  

RunwayFBU move: treat our proprietary exhaust as an asset, centralize it, then ship one LPreport agent and one dealhistory agent; swap UI tools as needed.