An essay on the venture distribution, why we believe the future is written in the tails, and what i⁶⁴ = 1 has to do with any of this.
Venture is not, statistically, a normal business. Its returns are not normally distributed — they are Pareto-distributed, with a power-law tail that means a handful of investments return more than all the others combined. Every operator in the asset class knows this; few build organizations that actually behave like it.
Most funds optimize for the mean. They build investment committees, deal teams, scoring rubrics. They aggregate signal. They survey the market. These mechanisms produce safer decisions, which is to say: decisions closer to the mean. They are excellent at avoiding catastrophe. They are also, structurally, incapable of taking the bets that the tail demands.
We don't think the bell curve is wrong, exactly — it's a fine description of past founders. The problem is that consensus prices a founder against the past. The companies that matter (Stripe in 2010, Anthropic in 2021, SpaceX in 2002) priced poorly against the consensus distribution. Each of them was four sigma off the mean of "what a fundable company looks like" — and each of them rewrote the distribution behind themselves.
Reality has a surprising amount of detail. Most founders smooth over the detail to make the pitch fit. The four-sigma founder is the one who can't. — internal memo, 04 Mar 2024
We've spent a career tuning a process to find them. It looks deceptively simple from the outside: read the deck, take the call, decide in 72 hours. The real machinery is the calibration — which signals are noise, which are tail-revealing.
Our name is a small mathematical joke that points at a larger discipline.
i is the imaginary unit — the canonical "doesn't make sense
yet" number. Raised to the fourth power it returns to 1; raised to the
sixty-fourth, it remains 1. Any imaginary, with enough iterations, becomes
real.
Substitute "founder" for "i" and the metaphor writes itself. The companies that look imaginary at pre-seed — the wrong school, the wrong sector, the wrong country, the wrong shape — are the ones that compound to 1, which is to say: become real. The math doesn't care that the founder is non-consensus. The math just requires you to hold long enough for the rotation to complete.
We're based in Monaco for the unsexy reasons. Capital density (per square kilometre, no other jurisdiction is close), tax treatment of carry, time zone, and a flight time to every major European hub under three hours. We did not move to Monaco for the marketing. We moved because the unit economics of running a small fund are dramatically better here, and the savings flow back to LPs.
We're not a "European fund." We invest where outliers are, which has meant a heavy European weighting, with US and rest-of-world opportunistically. We expect a similar mix for the fund.
We won't run a deal team. We won't write a partnership memo. We won't push a portfolio company into a round it doesn't need. We won't index. We won't co-invest into 200 companies a year and tell ourselves it's venture. We won't ask founders to sit through pitches with people they'll never speak to again.
— A. Mirenkov, C. Roux, Y. Hayashi · Monaco · February 2026.