Qanot / Fund mathematics
Every number a founder is never shown before signing.
Founders negotiate with venture funds without knowing what the fund needs in order to work. That asymmetry is not useful to anybody. Below is the whole arithmetic — the power law, the exit a fund requires, what portfolio construction is actually for, how ownership decays, and why the first six years look like failure. Two of the models are runnable.
The fund is its best company. Everything else is rounding.
In a normal distribution, the mean tells you about the population. In a power-law distribution it tells you almost nothing, because the extreme values dominate the sum. Venture returns are firmly the second kind.
The empirical shape, consistent across decades of fund data: roughly six in ten investments return less than the money put in, about three in ten return between one and five times, and a very thin slice — one or two percent — returns more than twenty times and pays for everything else.
The uncomfortable implication is that a fund's job is not to be right often. A fund that is right 40% of the time with no outlier loses money. A fund that is wrong 95% of the time and owns one company that returns 50× is a top-quartile fund.
This is why "de-risking" is the wrong frame at the formation stage. The variance is the asset. A company with no chance of failing also has no chance of being the outlier, and a portfolio built from those is arithmetically incapable of returning a venture fund.
What has to happen for a fund to return capital.
A fund promises its investors a multiple on committed capital. Work backwards from that promise and you get a single required number: the enterprise value of the best outcome. Here is the chain.
proceeds = fund_size × target_multiple
// the share of that carried by the single best company
top_proceeds = proceeds × concentration
// the enterprise value that produces it, given final ownership
required_exit = top_proceeds ÷ ownership_at_exit
Assumes 60% of committed capital deployed as initial cheques, the balance reserved for follow-on, and dilution already reflected in ownership-at-exit. Illustrative model, not a projection of returns.
Push the fund size slider and watch what happens. At $25M a single $640M outcome returns the fund three times. At $250M the same portfolio needs a $6.4B company — and there are perhaps a few dozen of those created worldwide in a good year. This is the entire argument for a small first fund in an emerging market: the required outcome has to be one that the market can actually produce.
Run the fund two thousand times.
Monte CarloPosition count is the only lever an early-stage fund genuinely controls. You cannot make a company succeed, but you can decide how many draws you take from the distribution. This simulator draws each position from the empirical outcome distribution above and sums the result, two thousand times over.
Move the position count and watch the left tail. A twelve-position fund has a meaningful chance of returning almost nothing — not because the manager is bad, but because twelve draws from a fat-tailed distribution frequently miss the tail entirely.
What the simulator is not. It assumes every position is an independent draw from a fixed distribution. Real funds have correlated outcomes (one macro cycle hits everything), selection skill that shifts the distribution, and reserves that concentrate capital into winners after the fact. It is a demonstration of why position count matters, not a forecast of any fund's return.
Everybody's slice gets smaller. That is the design.
Founders tend to experience dilution as loss. It is more usefully read as the price of enlarging the thing being divided. Giving up 20% of a company that is now worth ten times more is not a concession; it is the mechanism.
The table shows a typical path from a first cheque to exit. Note the last column: the first-cheque investor who bought 8% ends up with roughly 3% — which is precisely why the required-exit number in the model above uses ownership at exit and not ownership at entry.
The practical lesson for founders: the round that hurts is never the one with the largest percentage. It is the one taken at a price the business cannot grow into, because that is the round that forces a down round later — and a down round is where preferences and ratchets do real damage to common stock.
| Round | New shares | Founders | Qanot |
|---|---|---|---|
| First cheque | 8.0% | 83.0% | 8.0% |
| Seed | 18.0% | 68.1% | 6.6% |
| Series A | 20.0% | 54.4% | 5.2% |
| Series B | 15.0% | 46.3% | 4.5% |
| Series C | 12.0% | 40.7% | 3.9% |
| Option top-ups | 8.0% | 37.5% | 3.6% |
| At exit | — | 37.5% | 3.6% |
The money held back is the money that matters.
A first-cheque fund that deploys 100% of its capital into initial positions has guaranteed itself the dilution path in the table above. Reserves exist to buy that ownership back — selectively, and only in the companies that have already shown they work.
This is the one place where a fund gets to use information the market does not have yet. We have sat in the weekly revenue review for eighteen months. We know which company's growth is real. Reserves are how that knowledge converts into ownership.
- Reserve ratio
- Roughly 40% of the fund, held for follow-on. Deployed into approximately one third of the portfolio.
- Trigger
- A priced round led by an institutional investor we did not introduce. Third-party validation, not our own optimism.
- Objective
- Defend the entry position through Series A, and increase it where the evidence has improved rather than merely persisted.
- What we will not do
- Follow on to protect a mark. Writing a second cheque to avoid recognising that the first was wrong is the most expensive habit in this business.
The first six years look like failure.
A fund draws capital and pays fees from day one, and returns nothing for years. Plotted, the net position falls before it rises — the J-curve. A fund in year four that is showing a negative net multiple is not underperforming; it is on schedule.
This is worth understanding as a founder because it explains behaviour that otherwise looks irrational. A fund in the first half of its life is buying options and will tolerate ambiguity. A fund in year eight needs distributions and starts pushing for exits that may not serve you.
Ask any investor what vintage their current fund is. It tells you more about how they will behave on your board than anything on their website — including this one.
Six things this arithmetic should change about how you raise.
- Ask the fund size
- It tells you the exit they need. A $500M fund cannot be made whole by your $80M acquisition, and will steer you away from it regardless of whether it is the right outcome for you.
- Ask the vintage
- Year two and year eight are different investors wearing the same name.
- Ask the position count
- A fund with 15 positions needs you specifically to work. A fund with 60 is buying the distribution. Both are legitimate; they produce very different board behaviour.
- Do not optimise the headline price
- The valuation you cannot grow into within 24 months is a liability that compounds. Take the clean round.
- Read the preference stack, not the number
- A $1B valuation with a 2× participating preference can leave common stock worth less than a $400M clean exit.
- Default-alive beats a good story
- It is the only position that lets you decline a bad round — and the ability to decline is the whole of your negotiating power.