Investing in Early-Stage Companies

In a venture fund the median investment loses money, and that is the design rather than a failure of execution. Most companies return little or nothing, several return roughly what went in, and a small number produce nearly all of the gains. A manager who avoided the losses would also have missed the outsized winners — both come from the same population.

A venture fund raises committed capital, invests in young companies over its first several years, reserves a substantial portion for follow-on rounds, and waits. Investments are typically made in preferred stock, which sits ahead of founder and employee common stock and carries a liquidation preference plus governance rights — so a modest exit can return capital to the fund and nothing at all to the founders.

Liquidity arrives only at an acquisition or a public listing, which is why the horizon here is the longest in this book — a decade or more. Interim reported values come from the price of the most recent funding round, so a fund's marks can rise because someone else paid more and fall when the next round prices lower.

Venture outcomes follow a power law: a steep curve where a few results dwarf everything else, rather than a bell curve with a middle. The reason is asymmetric arithmetic — the worst an investment can do is lose what went in, and the best has no ceiling. Expected value therefore lives in the extreme right tail, and a strategy built on the tail has to be structured to reach it.

Diversification here is a mathematical requirement rather than a preference. Concentration does not add variance to expected return the way it does in public equities; it removes access to the only outcomes that make the asset class work. Spreading capital across many companies and several funds is the strategy, not a refinement of it.

The same math changes how you read a fund's early reports. A fund three years in with most positions flat or written down and one marked up sharply is behaving as the model predicts. A fund where nearly every position is marked modestly higher is the more concerning picture — it suggests conservative marking, or a portfolio without a genuine outlier.

Manager access matters more in venture than in any other category in this book, and the reason is structural. Anyone can buy the same public share at the same price; in venture, the funds with the strongest histories are oversubscribed and ration capacity by relationship rather than price. You cannot pay your way into a closed fund, so what is available to you is a different question from what is good.

That creates adverse selection: the venture funds most easily available to an individual are, on average, the ones with the least demand from institutions that have evaluated this asset class for decades. That is not a reason to avoid venture. The question it raises is whether a fund's availability reflects genuine capacity, a deliberate strategy of admitting individuals, or the absence of other buyers.

• Fund of funds — one commitment spread across many underlying venture funds, which delivers the diversification the math requires and adds a second layer of fees.
• Feeder or platform vehicles — pooled access to a single named fund at a lower minimum, with an added fee layer and no governance rights.
• Emerging managers — newer funds with genuine capacity, where access is real and the track record you would want does not yet exist.
• Direct company investments and syndicates — the most concentrated route, and the one where the power law is least forgiving of too few positions.

Picture a venture fund that backs twenty early-stage companies. Over a decade, perhaps half return little or nothing, several return roughly the capital invested, a few do genuinely well — and one becomes a breakout that returns many times what went into it. That one company can carry the fund while the majority of its holdings disappoint.

The discipline is accepting the losers as the cost of being positioned for the rare winner — sound only when capital is spread across enough companies for an outlier to appear. Concentrate into two or three names and you take the full loss profile of the asset class while cutting your odds of holding the outcome that pays for it. An investor who needs steady results, predictable cash flow, or the money back within a decade is in the wrong asset regardless of the fund.

Venture is equity risk in its most concentrated form, so it belongs in the growth portion of a portfolio and should be funded from equity rather than from something providing stability. The allocation should be small enough that losing all of it changes nothing important about your plan; if that does not hold, no amount of manager quality makes the position appropriate — the loss scenario here is normal, not a tail case.

The failure mode I see most often is not picking a bad fund — it is writing one large check into one fund in one year, stacking three concentration risks at once: a single manager, a single vintage, and too few underlying companies. The remedy is pacing — smaller commitments across several years and several managers, which spreads you across market environments as well as companies. Sizing and liquidity works through the mechanics.

One tax note belongs before you invest rather than after: equity in certain early-stage companies can qualify for favorable capital gains treatment under the qualified small business stock rules, and qualification depends on facts about the company and your holding period, fixed at the time of investment. This matters more for direct investments and syndicates than for fund interests, and the rules belong in a conversation with your CPA — see tax mitigation.

Venture is the only asset class in this book where being wrong most of the time is compatible with a good result — that inversion is why it defeats intuitions built on every other kind of investing. Judge it on breadth, on access, and on whether a complete loss would be merely disappointing — not on the proportion of investments that worked.