Venture Returns Are Decided at the Sale, Not the Entry, and the 10-Year Fund Forces the Wrong One, HarbourVest's Voss Argues

53 minutes ago

4 min read

What's New

In a power-law asset class, the decision to sell a winner is as consequential as the decision to buy it, and the 10-year closed-end fund is a mechanical trigger that pushes managers to sell the one company that carries the fund. Scott Voss, Partner at HarbourVest, argues this in a podcast on How I Invest. A theoretical portfolio of 15,000 companies built from HarbourVest and third-party data shows one company returning 10% of total value and the top 10% returning 90%. When a fund reaches its term and no natural exit exists, Voss says the manager should create one through the secondary market, or extend the fund rather than cut the winner. Investors should judge managers on divestment discipline, not only origination.

Why It Matters

The standard LP playbook asks managers to take chips off the table at 1x to 3x basis and let the rest ride. Voss shows the cost of that instinct with Klarna, where investors sold into a $50 billion private round ahead of a $15 billion IPO and a sub-$10 billion market today, a 10x call that cuts both ways. On the other side are LPs who treat venture like buyout and cannot stomach loss ratios, and managers who monetize early and lose their edge. HarbourVest runs $160 billion across primaries, secondaries, and directs and sells access to the funds Voss describes, so the structure-agnostic thesis is also his product.

Big Picture Drivers

The power law is the business model: One company in 15,000 returned 10% of value; 10 returned a third; the top 10% returned 90% of a 3x portfolio on $15 billion of cost. Every great portfolio carries more losses than average.
Fund terms create artificial exits: Most venture funds are 10-year drawdown vehicles. Voss cites managers who went to LPs in years 13 to 15 and extended to 25 years, and a Roblox investor who rolled all earned economics into continuation vehicles to stay aligned and hungry.
Consensus starts at $10 billion: Before SpaceX went public, one private company exceeded $1 trillion, roughly 10 sat between $100 billion and $1 trillion, and about 100 between $10 billion and $100 billion. That 100 to 120 is the consensus everyone chases. The money round is picking which of 1,500 unicorns will graduate into it.
Access is structure-agnostic: The goal is economic interest in the short list that matters, via a primary fund at 2.5 and 25, a co-invest that may be fee-free, or a secondary where the seller pays no fee. The same asset can be a 100x, a 3x, and a 2x depending on entry point.
Partner attribution, not firm brand: HarbourVest maps rainmakers, emerging partners, and laggards inside each firm, while acknowledging a laggard's portfolio can inflect years after they were let go. Thesis timing, not skill, explains many fintech and crypto records.
Firms lose their touch through strategy drift or leadership: Investing where the market is not going, or founders failing to hand over to the next generation, explains most declines. Wealth and lost fire in the belly explain the rest.

By The Numbers

10% to 90%: Share of companies that produced share of value in HarbourVest's 15,000-company, two-decade dataset.
$50 billion to under $10 billion: Klarna's last private round versus its current public value, with a $15 billion IPO in between.
25 years: Life of at least one venture fund after a conscious extension by a manager holding a generational asset.
$104 billion: Meta's IPO valuation, the record for a US listing for 13 years until SpaceX arrived at 15 to 20 times that size.
50 to 100 basis points: Annual fee of a multi-manager venture fund, with 0 to 5% carry, versus 2.5 and 25 or more on a direct commitment.
$25 billion: Capital HarbourVest raises and invests a year across private markets, with venture around 15%.

Key Trends to Watch

Continuation vehicles as the default for fund-life winners: With roughly $3 trillion said to sit in funds past their 10-year term, watch how many managers roll their power-law companies forward versus selling to secondaries buyers who then own the upside.
Insider-led tenders as a price signal: SpaceX ran tenders every six months with much of the liquidity coming from insiders. Watch whether other mega-privates adopt the same pattern and whether insiders keep buying.
Scale venture becoming an oligopoly: Voss expects about 10 firms to dominate venture at scale within a decade while zero-to-one investing stays fragmented. Watch fund sizes at the top and whether $50 to $100 billion private raises become routine.
The definition of venture stretching: A company with $30 billion in revenue growing 300% and raising $50 to $100 billion privately is still called venture. Expect the label, and the benchmarks, to be revisited.

Memorable Quotes

"It's everything. It's the business model." Voss on the power law, and why LPs who think in buyout terms misjudge venture.
"The ones that are like super consequential, I can count on two hands." Thousands of decisions in 27 years, and only a handful, mostly around when to hold or sell, that moved outcomes.
"A good term sheet does not make a bad deal good and a bad term sheet does not make a good deal bad." The rule he learned from the partner who trained him, applied to fee arguments across entry structures.
"Convince yourself you should be in it for the next 20 years. And don't change your mind." His advice to anyone entering venture now, when the best vintages follow the worst.

The Wrap

The thesis holds if managers who extended or rolled their power-law winners through secondaries and continuation vehicles outperform peers who took chips off the table at term, and if structure-agnostic entry via co-invest and secondaries delivers top-quartile exposure at lower blended fees than primary commitments alone. It fails if the current run of trillion-dollar privates reverts, making the Klarna-style early sale the smarter call across the cohort, or if continuation vehicles become a way to defer marks rather than capture upside. Voss says dispersion in venture collapses over 15-year windows and widens over single vintages. The 2020 to 2022 funds hitting year 10 between now and 2032 will show which sell decisions were the consequential ones.