Welcome back to the EUVC Podcast, your inside track on the people, models, and math reshaping European venture.
This week, Andreas talks with Damian Cristian and Guy Conway, co-founders of Rule 30 - an AI research lab building what they claim is the world’s first fully systematic venture strategy. We go deep on the difference between “data-driven” (hygiene) and decision-driven (engine), why labels matter, and how portfolio math crushes intuition.
They unpack founder-trajectory signals, graph-based network evolution, market topology (yes, biology-inspired stats), and a portfolio design targeting 3x+ minimum returns with 97.5% confidence. We also debate the “access myth,” party rounds, and why they won’t sell their alpha.
Whether you’re an LP testing managers, a GP rethinking reserves, or a founder curious how algorithms “see” you - this one’s for the nerds and the pragmatists.
Here’s what’s covered:
01:46 | What is “Quant VC” and how it differs from traditional venture
06:39 | Why pre-seed isn’t an access problem — it’s a triage problem
09:55 | Can AI really make investment decisions at pre-seed?
14:13 | Training the model on 15 years of startup data to find top-decile winners
20:55 | The “Outlier Trajectory” of founders — decoding team evolution through data
26:42 | Why Rule 30 calls itself an AI Research Lab, not a VC fund
35:36 | Portfolio construction math: the danger of the “middle” strategy
55:57 | Follow-ons vs upfront bets — why they avoid reserves entirely
61:40 | Access myth-busting — why 99 % of pre-seed deals are open to smart capital
🎧 Listen on Apple or Spotify — chapters are set for easy navigation.
✍️ Show Notes
Quant vs Data-Driven (and why it matters)
“Data-driven” is hygiene (more signals, cleaner CRMs). Quant is a decision engine that actually picks based on learned patterns — and sticks to them even when the room culture disagrees.
Training set: yearly cohorts since 2010; outcome label = top decile in valuation delta from first round (portfolio-level DPI proxy), not “pick unicorns.”
Signals that Consistently Pop
Founder Outlier Trajectory: time-series of roles, network slope, and cohort-relative progression; compares “ideal team” vs actual team and measures delta against 20k+ labeled comps.
Graph evolution: pre-investment network dynamics that foreshadow which investors are likely to show up.
Market topology: abstract competitive spaces with stats inspired by quant finance & biology.
“People assume there isn’t enough data at pre-seed. The truth is there isn’t enough human-computable data. The algorithms can.” — Damian
Portfolio Construction (a quantified stance)
Two strategies work historically: very wide (index-like) or ultra-concentrated (true benchmark-style). The “30–40 deals + vibes + reserves” middle is a valley of death.
Rule 30 targets 75–85 initial checks, no reserves, and separate follow-on vehicles only for strategic reasons.
Goal: reduce volatility of the asset class; design for ≥3x with high confidence, not lottery-ticket 40x paired with 0.3x.
“If you can write the bigger check upfront, EV says do it. ‘Double-down later’ sounds great — it’s usually worse than sizing right at entry.” — Damian
Access at Pre-Seed (myth-busting)
Hypothesis borne out so far: access isn’t the bottleneck for ~99% of pre-seed rounds; the 1% you can’t access are usually mispriced anyway.
Technical founders like engaging an algorithmic IC; auto-memos increase trust and speed.
Why an AI Research Lab?
They won’t sell the alpha as SaaS. Mandate is to mine it — possibly with LP partnerships — and expand from pre-seed to later stages and adjacent private-market strategies.
