Power Law
Quick Definition
A power law is a statistical distribution where a small number of outcomes produce the majority of total results. In a normal distribution, most outcomes cluster around the average. In a power law distribution, the average is almost meaningless because a tiny fraction of outcomes dominates everything else. Venture capital returns follow a power law: a handful of blockbuster investments generate more return than all other investments combined.
What It Means
Most people think about risk and return using a normal distribution, the familiar bell curve. In a normal distribution, most outcomes cluster near the mean, and extreme outcomes are rare. The average is a useful summary statistic. You can plan around it.
Power law distributions work differently. A small number of observations account for the majority of the total. The distribution has no natural ceiling. The gap between the best outcome and the median outcome is enormous. The average is pulled by the outliers and tells you almost nothing about what to expect from any single observation.
This has profound implications for investing. According to iCapital research published in May 2026, just 3% of venture capital deals account for nearly half of the industry's total returns. That single fact should change how investors think about the asset class. It means that in venture capital, how you allocate matters more than whether you allocate at all. Fund size, stage focus, ownership economics, and manager selection all determine whether a portfolio captures a few exceptional winners or spreads exposure across average outcomes.
The power law is not unique to venture capital. It appears in book sales, movie box office results, scientific citation counts, city populations, website traffic, and wealth distribution. Anywhere that cumulative advantage, network effects, or winner-take-most dynamics operate, power law distributions tend to emerge.
How It Works
The Mathematics
A power law distribution has the form P(x) proportional to x raised to the power of negative alpha, where alpha is the scaling exponent. When alpha is small (between 1 and 2), the distribution has a very heavy tail, meaning extreme outcomes are common enough to dominate the total. When alpha is larger, the distribution looks more like a normal distribution with thinner tails.
For venture capital, the effective alpha is low enough that the tail dominates. This means that the expected value of a single investment is driven almost entirely by the probability of an outlier outcome, not by the median or most likely outcome.
The Venture Capital Data
Two of the most-cited studies on venture return distribution come from Horsley Bridge and Correlation Ventures, and they agree on the shape even though they used different datasets.
Horsley Bridge, a longtime LP across dozens of US venture funds, analyzed 7,000 of its own portfolio investments made from 1975 through 2014 and found that just 6% of those deals generated 60% of total returns. A fund's success is decided by a small handful of positions, not its average deal.
Correlation Ventures studied more than 21,000 financings from 2004 to 2013 and then expanded to over 27,000 from 2009 to 2018. They found that 65% of deals returned less than the capital invested. Only about 4% returned more than 10x, and a mere 0.4%, fewer than 1 in 200 investments, returned more than 50x. Those sub-1% outcomes are where the bulk of aggregate industry returns actually live.
| Return Multiple | % of VC Deals | Share of Total Returns |
|---|---|---|
| Less than 1x (capital loss) | 65% | Minimal |
| 1x to 5x | ~25% | Moderate |
| 5x to 10x | ~6% | Significant |
| 10x to 50x | ~3.6% | Large |
| 50x+ | 0.4% | Dominant |
AI Is Steepening the Power Law
A 2026 analysis from Commonfund titled "AI is Steepening the Power Law in Venture Capital" found that AI is concentrating venture returns even further. Looking at venture exit events since 2023, the top 1% of companies represent 80% of total venture capital exit value (or 45% excluding the $1.77 trillion SpaceX IPO), up from 17% from 2005 to 2010 and 34% from 2017 to 2022.
AI now absorbs more than 80% of US venture dollars by early 2026, up from 64% in H1 2025. A few companies are attracting the majority of investment dollars and commanding the steepest valuations: OpenAI was recently valued at $852 billion, Anthropic at $965 billion, and SpaceX at $1.25 trillion while private. The average top-five private company was worth about $25 billion in 2015 and $35 billion in 2020, then jumped to $473 billion as of June 2026, roughly 17 times the 2015 level.
The winners are also arriving faster. Stripe data shows AI companies reaching $30 million in annualized revenue in a median of 20 months, versus five-plus years for the prior software generation. At the top, revenue scales even faster: Anthropic reported a $47 billion run rate in late May 2026.
Implications for Portfolio Construction
Power law distributions require a fundamentally different approach to portfolio construction than normal distributions.
| Strategy | Normal Distribution | Power Law Distribution |
|---|---|---|
| Goal | Maximize average outcome | Maximize exposure to outliers |
| Diversification | Moderate number of positions | Many positions to capture rare winners |
| Position sizing | Equal weight or risk-weighted | Concentrate in highest-conviction bets |
| Risk management | Limit downside on each position | Accept frequent small losses |
| Evaluation | Track average returns | Track hit rate on outliers |
| Key metric | Mean return | Tail outcomes, maximum return |
In a power law world, you are not trying to maximize the expected value of the average deal. You are trying to maximize your exposure to the right tail. A model optimized for "most likely to succeed" will systematically strip out the high-variance, non-consensus bets, because those are exactly the ones that look riskiest on the mean. Let that model gate your pipeline and it will hand you a portfolio of sensible companies that never returns a fund.
Fund Size and Power Law Capture
iCapital's 2026 research found that sub-$350 million funds have meaningfully outperformed larger vehicles on both internal rate of return (IRR) and total value to paid-in capital (TVPI). Larger funds provide exposure to scaled AI leaders and late-stage opportunities but face mathematical constraints that compress return multiples. The reason is structural: a $10 billion fund cannot make meaningful investments in early-stage companies without owning the entire company, so it must deploy in later rounds where valuations are higher and return multiples are lower. The power law outcomes that make venture capital attractive happen in early-stage investments, and large funds are structurally excluded from capturing them at scale.
Real-World Examples
The WhatsApp Investment
Sequoia Capital invested approximately $60 million in WhatsApp across multiple rounds, the only institutional investor in the company. When Facebook acquired WhatsApp for $19 billion in 2014, Sequoia's stake was worth approximately $3 billion, a 50x return. That single investment returned more than many venture funds raise in total. This is the power law in action: one position can define a fund's entire track record.
The SpaceX Outcome
SpaceX reached a private valuation of $1.25 trillion before its IPO, making it one of the largest wealth creation events in venture capital history. Founders Fund's early investment in SpaceX, reportedly around $20 million, became worth billions. The Commonfund analysis noted that excluding the $1.77 trillion SpaceX IPO, the top 1% of companies still represent 45% of total venture exit value since 2023, an extraordinary concentration.
The AI Concentration
The current AI cycle is steepening the power law to an unprecedented degree. OpenAI, Anthropic, and a handful of other AI companies are absorbing the majority of venture capital and generating the largest potential outcomes. Companies that would have gone public at $25 billion in 2015 are staying private at $500 billion or more in 2026. This means the power law outcomes are concentrating in fewer companies, and the gap between winners and everyone else is widening.
What This Means for Individual Investors
Most individual investors do not have access to top-tier venture funds. But the power law principle applies beyond venture capital. In public equity portfolios, a small number of positions often drive the majority of returns. Research shows that across 20-year periods, less than 5% of public stocks account for all the net wealth creation in the stock market, when measured against Treasury bills. The rest collectively match or underperform cash. This means that diversification is important not because every position will be a winner, but because you need enough at-bats to capture the few that are.
Key Points to Remember
- A power law distribution is one where a small number of outcomes account for the majority of total results
- In venture capital, 3% of deals generate nearly half of all returns, and 65% of deals lose money
- AI is steepening the power law: the top 1% of companies now represent 80% of VC exit value since 2023
- Power law investing requires maximizing exposure to outliers, not optimizing average outcomes
- Sub-$350M venture funds outperform larger funds because they can make concentrated early-stage bets
- The power law applies to public equities too: a small fraction of stocks drives most of the market's wealth creation
- Understanding power laws changes how you think about risk, diversification, and portfolio construction
Common Mistakes to Avoid
- Evaluating venture investments by average returns: The average venture investment loses money. The median venture investment returns less than 1x. Evaluating the asset class by its average misses the point entirely. The relevant question is whether your portfolio captures outliers, not whether your average deal performs well.
- Under-diversifying in power law contexts: Because most investments fail, you need many positions to have a reasonable probability of capturing an outlier. A portfolio of 5 venture investments has a high probability of losing money. A portfolio of 30 to 50 positions has a much better chance of including a winner that returns the fund.
- Over-diversifying and diluting outlier exposure: While you need enough positions to capture outliers, too many positions dilute your ownership in the winners. If your best investment returns 100x but you owned 0.1% of it, the impact on your portfolio is small. The art is finding the right balance between enough at-bats and meaningful ownership in the winners.
- Ignoring survivorship bias: Studies of venture returns often suffer from survivorship bias. Funds that performed poorly may not raise subsequent funds and disappear from databases. The reported average returns of surviving funds overstate the experience of the typical investor.
- Applying normal-distribution thinking to power law domains: If you evaluate a venture investment by asking "what is the most likely outcome," you will reject the high-variance bets that drive power law returns. The most likely outcome for any early-stage startup is failure. The relevant question is "what is the upside if this works, and how much can I own if it does."
Related Concepts
The power law distribution is central to understanding risk and return in venture capital and other winner-take-most markets. It explains why diversification works differently in power law contexts than in normal-distribution contexts. Behavioral finance helps explain why investors systematically misjudge power law opportunities, overweighting the probability of failure and underweighting the magnitude of potential success. Survivorship bias distorts our perception of power law returns by hiding the failures. For angel investors, understanding the power law is the single most important concept for portfolio construction. For practical investing guidance, read our article on common investing mistakes beginners make. You can also use our investment return calculator to model how different return distributions affect portfolio outcomes. For current data on venture capital power law dynamics, the iCapital research on venture capital in the AI era provides 2026 analysis.
Frequently Asked Questions
Q: Does the power law apply to public stock investing? A: Yes, though less extreme than in venture capital. Research by Hendrik Bessembinder shows that across 20-year periods, less than 5% of public stocks account for all the net wealth creation in the stock market. The majority of stocks collectively underperform Treasury bills. This means that broad index funds work partly because they capture the few outlier stocks that drive most returns, not because every stock in the index is a good investment.
Q: How many investments do I need to capture a power law outcome? A: In venture capital, most practitioners recommend 30 to 50 portfolio company investments to have a reasonable probability of capturing an outlier. With fewer than 20 investments, the probability of missing the winner that returns the fund is high. The exact number depends on your stage focus, check size, and conviction in individual deals.
Q: Is the power law getting steeper or flatter over time? A: The evidence suggests it is getting steeper. The Commonfund 2026 analysis found that the top 1% of companies now represent 80% of VC exit value since 2023, up from 17% in 2005 to 2010. AI is accelerating this trend by concentrating capital and outcomes in fewer companies. This means capturing outliers is both more important and more difficult than in previous cycles.
Q: Can I invest in power law outcomes without investing in venture capital? A: To some extent, yes. Public equity markets also exhibit power law dynamics, though less extreme. Broad index funds capture the outlier public stocks that drive market returns. Concentrated portfolios of high-conviction public stocks can also capture power law outcomes, but with higher risk of missing the winners. Options strategies, particularly long-dated call options on high-conviction stocks, can create power law-like payoff profiles with defined downside.
