Venture Power-Law Underwriting Model
Underwrite each deal for extreme upside and the portfolio for frequent failure
- Difficulty
- Advanced
- Time to result
- ~ongoing to results
- Steps
- 6
- Confidence
- 91%
The model begins with the observation that venture outcomes follow a power law: many startups fail, a meaningful group produces acceptable results, and a small minority creates nearly all portfolio alpha. Investors therefore assess each early-stage opportunity against a very high potential return threshold while constructing the overall fund to survive frequent losses. The transcript offers illustrative targets of one hundred times invested capital at seed and ten times at Series A, alongside a portfolio assumption of forty percent failure, fifty percent success, and ten percent super-success. These figures are not universal forecasts; they express the mechanism. A viable investment must have enough plausible upside to matter at fund level, and the portfolio must contain enough independent attempts for rare winners to emerge.
Origin
Extracted from Masters in Business, where Sarria Darabi explained TMV's early-stage underwriting assumptions and the concentration of alpha among rare winners.
Core principles
- 01A small minority of investments generate most fund returns
- 02Early-stage portfolios must expect substantial failure
- 03Required upside should reflect investment stage
- 04Fund economics matter more than the average company outcome
- 05Better sourcing does not eliminate power-law risk
How to run it
- 1
Define the Fund Objective
Set the fund-level return required by its mandate, duration, fees, and risk. Translate that objective into the amount of value the portfolio must ultimately create.
Pro tip Model returns in both gross and net terms.
Watch out An attractive company outcome may still be immaterial to a large fund.
- 2
Set Stage-Specific Upside Thresholds
Assign higher required return potential to earlier, riskier rounds and lower—but still substantial—targets to later stages.
Pro tip Use return targets as screening hurdles rather than precise predictions.
Watch out A large theoretical market does not prove that a company can capture enough value.
- 3
Model Outcome Buckets
Estimate how much of the portfolio may fail, succeed modestly, or become a super-success. Test several distributions rather than trusting one point estimate.
Pro tip Include dilution, follow-on capital, and ownership changes in the model.
Watch out Historical averages may not fit a specialized strategy or current market cycle.
- 4
Test Fund-Returning Potential
For each candidate, calculate what ownership and exit value would be needed to materially affect the fund. Reject deals whose best plausible outcome remains too small.
Pro tip Work backward from a fund-returning outcome to required revenue, market share, and exit value.
Watch out Do not turn an extreme upside case into the base case.
- 5
Diversify Independent Bets
Build enough exposure across companies, founders, markets, and vintages to avoid relying on one forecast. Preserve sufficient ownership in the strongest performers where appropriate.
Pro tip Track concentration by underlying risk driver, not merely company count.
Watch out Many startups exposed to the same market shock do not constitute true diversification.
- 6
Recalibrate with Evidence
Compare actual failures, exits, markups, and dilution with the original portfolio assumptions. Adjust future pacing and underwriting thresholds when evidence consistently diverges.
Pro tip Separate temporary valuation marks from realized performance.
Watch out Do not rewrite assumptions solely to justify prior decisions.
In the wild
A seed fund models eight failures, ten ordinary successes, and two super-successes. It invests only where a credible extreme-upside path exists, recognizing that the two largest outcomes may need to create most of the fund's value.
→ The portfolio is designed to tolerate common losses while preserving exposure to rare, fund-driving winners.
Common mistakes
Expecting Every Deal to Win
A portfolio that assumes uniformly positive outcomes misunderstands early-stage uncertainty and is likely to overpay or under-diversify.
Using Headline Multiples Literally
Targets such as 100x at seed are underwriting lenses, not promises. Probability, dilution, ownership, and exit feasibility still require explicit analysis.
Ignoring Fund Size
An exit can be successful for a founder but too small to affect the investor's fund-level return.
Is it for you?
Best for
It is best for seed and Series A investors constructing diversified portfolios of high-risk, high-upside companies.
Not ideal for
It is not ideal for capital-preservation mandates, concentrated portfolios, or investors requiring predictable cash flows.
From the transcript
“traditionally BC's underwrite one hundred x for a seed investment, ten x for a series A investment.”
“You're also underwriting for a forty percent fail rate, fifty percent success rate, and ten percent super success rate.”
“And it's those ten percent of companies that really deliver all of the alpha for any given fund, not just mine.”
From the episode
At the Money: Finding Overlooked Private Investments