Venture capital performance is among the most discussed topics in tech industry coverage but among the most inconsistently documented. Headlines about specific successful investments dominate public attention. The actual performance distribution receives substantially less examination.
This analysis uses public data sources to examine actual VC returns across the past decade. The objective is documented evidence about how venture capital has actually performed rather than the dominant narratives.
Methodology
The analysis draws on:
Cambridge Associates Venture Capital Index data (industry-standard VC return tracking).
Pitchbook data on specific fund performance (where publicly available).
Public information from limited partner disclosures (university endowments, public pension funds with disclosure requirements).
Academic research on VC returns including Korteweg, Sorensen, and others.
SEC filings from publicly-traded VC firms and parent organizations.
The data sources have known limitations including selection bias (poor performers underreport), reporting timing (returns develop across long periods), and methodology differences across sources. The analysis acknowledges these limitations.
Aggregate venture capital performance
Across the past decade, aggregate VC performance shows specific patterns:
Industry returns. The aggregate VC industry has produced returns approximately matching public market returns over recent decades, with substantial volatility and time-period dependence.
Top quartile performance. Top quartile funds have produced returns substantially exceeding public markets. The outperformance is real and has been sustained.
Median performance. Median fund performance has been roughly comparable to public markets. The median fund hasn't justified the illiquidity premium VC investments theoretically deserve.
Bottom quartile performance. Bottom quartile funds have substantially underperformed public markets. The underperformance is consistent across periods.
The distribution is highly skewed. Aggregate returns reflect a small number of substantial successes alongside substantial failures.
The fund-level distribution
VC fund returns show specific distribution patterns:
Approximately 50-60% of funds return less than 1.5x invested capital across their lifetime.
Approximately 25-30% return 1.5x-3x invested capital.
Approximately 10-15% return 3x-5x invested capital.
Approximately 3-5% return more than 5x invested capital.
The skewed distribution means most funds produce modest returns while substantial returns are concentrated in few funds.
For investors selecting between funds, the distribution implies most fund selections will produce modest returns regardless of selection criteria. The substantial returns require selecting from the few funds in the upper tail of the distribution.
Persistence of fund performance
Whether top fund performance persists across funds is contested:
Some research (particularly older research) suggested substantial persistence. Top funds tend to produce subsequent top funds.
More recent research (Sorensen and others) suggests persistence is weaker than earlier studies indicated when adjusting for various biases.
The contemporary picture is mixed. Some persistence exists but selecting future top funds remains difficult.
For investors, the persistence question matters substantially. If top funds reliably produce future top funds, investing with them is straightforward. If persistence is weak, investment becomes more challenging.
The contemporary evidence supports modest persistence with substantial uncertainty about specific future performance.
Returns by stage and sector
Returns vary substantially across investment stages and sectors:
Seed-stage investments. High variance. Substantial individual fund variability. Aggregate returns roughly comparable to other stages with more variation.
Series A. Highest aggregate returns historically. Concentrated investment by top firms. Substantial selection effects.
Series B and later. Lower variance. Lower aggregate returns. Different risk-return profile from earlier stages.
Growth stage. Different risk-return profile. Closer to public market returns with less variance.
Sector variations are substantial:
Software has consistently produced top sector returns over the period.
Hardware has variable returns with substantial cycle dependence.
Biotech has different risk-return profile than tech VC.
Specific subsectors (AI, fintech, etc.) have time-varying performance.
The variations affect which fund types are appropriate for which investor objectives.
The 2020-2024 period specifically
The 2020-2024 period has specific characteristics affecting VC returns:
2020-2021 saw substantial valuation increases and rapid fundraising. Many funds raised at peak valuations.
2022-2024 saw substantial valuation corrections affecting unrealized portfolio values.
Liquidity events (IPOs, acquisitions) declined substantially in 2022-2024.
Fund performance during this period is still developing. Many funds have substantial unrealized investments whose ultimate realization may differ from current marks.
Realistic expectations about 2020-2024 vintage funds: lower returns than the 2010-2014 vintage funds, with substantial uncertainty.
The period reminds analysts that VC returns develop across decade-plus time horizons. Mid-period assessment is necessarily preliminary.
What the data reveals about narrative claims
Several common narrative claims are not well-supported by the data:
Claim: VC consistently outperforms public markets. Aggregate VC has roughly matched public markets. Outperformance is concentrated in top fund tail.
Claim: VC is a reliable source of substantial returns. Median VC investment has produced modest returns. Substantial returns require unusual fund selection.
Claim: Top VC firms reliably maintain top performance. Persistence exists but is weaker than common claims suggest.
Claim: VC returns are uncorrelated with public markets. Substantial correlation exists, particularly during stress periods. Diversification benefits are real but smaller than sometimes claimed.
Claim: Specific sectors guarantee strong returns. Sector performance varies. Past sector winners aren't guaranteed future winners.
The popular narrative frequently overstates VC return characteristics in ways that don't match documented data.
The implications for limited partners
For limited partners (the institutions and individuals investing in VC funds), the analysis suggests:
VC requires substantial fund selection capability. The skewed return distribution means fund selection matters substantially.
Most LPs without substantial selection capability should expect aggregate-level returns. Aggregate returns may not justify VC's illiquidity and complexity premium.
Top-quartile fund access is competitive and limited. LPs without specific access patterns may not be able to invest in top funds regardless of intent.
VC allocation should be sized appropriately for specific institutional capability. Larger institutions with substantial VC programs and dedicated staff have different optimal allocations than smaller institutions.
Performance assessment requires long time horizons. Mid-fund evaluation may not predict ultimate performance.
The implications for entrepreneurs
For founders considering VC funding, the data suggests:
Most VC-funded companies don't produce VC returns. The exit requirements for VC returns are substantial.
VC pressure for substantial growth reflects fund-level economics. Specific growth requirements may not match individual company optimal trajectories.
Alternative funding (bootstrapping, revenue-based financing, alternative equity structures) may produce better outcomes for specific situations.
VC selection matters for founders too. Working with specific funds has implications beyond the capital received.
The fund returns data informs founder decisions about whether to pursue VC and how to navigate VC relationships.
The implications for the broader ecosystem
For broader tech ecosystem analysis:
VC capital availability cycles substantially. The 2020-2021 boom and 2022-2024 correction illustrate the cyclical nature.
The cycles affect which companies get funded and which don't. Specific company outcomes depend partly on when they raised capital.
Aggregate VC influence on tech ecosystem depends on capital deployment patterns that have varied substantially across periods.
The tech ecosystem includes substantial non-VC-funded activity. The VC-focused coverage emphasizes a specific funding source that doesn't represent all of tech.
Analysis of tech ecosystem dynamics benefits from understanding VC as one specific element rather than the dominant element it sometimes appears.
What the data can't tell us
Limitations of the available data:
Selection bias affects available data. Funds with poor performance may not report consistently.
The decade window may have specific period effects that don't generalize.
Specific firm-level data is limited by what funds choose to disclose.
Cross-fund attribution is difficult. Specific successful investments often involve multiple funds.
Causal claims about what produces fund success are limited by the nature of available data.
The analysis identifies patterns rather than producing predictive models of specific fund performance.
What better VC data would enable
Improvements to VC data infrastructure could include:
Standardized reporting requirements for institutional funds.
Better integration with public markets data for benchmarking.
Longer time-horizon tracking for fund performance.
Better disclosure of specific portfolio company outcomes.
Improved methodology for cross-fund attribution.
The improvements would benefit LPs, regulators, and broader ecosystem analysis. They face resistance from VC firms with interests in limited disclosure.
The relationship between VC returns and tech outcomes
An important distinction: VC returns are not equivalent to tech industry health.
VC returns measure investor outcomes specifically. The metric reflects fund economics rather than broader tech ecosystem developments.
Tech industry health includes various dimensions (innovation, employment, consumer benefit, social impact) not captured in VC returns.
Specific tech developments may produce substantial value while producing modest VC returns. Other tech developments may produce substantial VC returns while producing limited broader value.
Analysis of tech industry health benefits from metrics beyond VC returns.
The conflation of VC performance with tech health is a common pattern that the underlying distinction can correct.
Conclusions
Venture capital performance across the past decade shows specific patterns. Aggregate returns roughly match public markets. Top quartile substantially outperforms. Bottom quartile substantially underperforms. The distribution is skewed.
For limited partners, the patterns affect optimal investment strategy. Most LPs benefit from understanding aggregate VC returns rather than expecting top-quartile performance.
For entrepreneurs, the data informs decisions about VC funding versus alternatives.
For broader tech ecosystem analysis, VC returns are one specific metric rather than equivalent to tech industry health.
The popular narrative about VC overstates returns in specific ways. Better information enables better decisions across all participants.
Recognition of what the data shows produces better-informed engagement with venture capital and the broader tech ecosystem it influences.
Citation
Berg, N. (2024). "Venture Capital Returns: A Decade of Documented Performance Data." Follow the Geeks VC Analysis Series.
