Founders raising venture capital tend to hear the same names over and over: Sequoia, Andreessen Horowitz and Benchmark. In Ilya Strebulaev’s telling, the more important number is this one: about 5% of venture investors created roughly 90% of the industry’s profits. The real question for a founder is not whether an investor is famous, but whether that investor belongs to that top slice.
The article is a guest piece by Strebulaev, a professor at Stanford Graduate School of Business, where he teaches venture capital and private equity and leads the school’s venture capital initiative. Blake Jackson worked with him on the ranking.
A data-based alternative to black-box VC lists
Strebulaev argues that the venture industry has lacked a transparent, fully data-driven way to answer that question. He points to the Forbes Midas List as the market’s common reference point, but describes it as largely opaque. When his team tried to reverse-engineer the public methodology, even their best-fit replication missed 49 of Midas’ own top 100 names. Among investors who appeared on both lists, the correlation was only about 0.27.
That gap led to the 2026 Strebulaev-Jackson VC ranking. According to the article, it is built from more than 230,000 investments made over nearly 30 years by close to 13,000 venture investors across more than 5,000 companies. Every score can be traced back to a specific investment in a specific company on a specific date. The methodology uses no editorial judgment, no manual adjustments and no company-submitted self-reporting, the article says.
The piece includes the full top 100, the methodology behind it and guidance for founders on what to do with a ranking once they have one.
What the score is meant to measure
Six factors drive the ranking, and Strebulaev frames them as behaviors founders should want in an investor.
The first is dilution. A 10% stake in the first round does not stay at 10% through exit if a company raises repeatedly. The model tracks how each investment’s ownership changes through later rounds.
The second is net profit. Turning $10 million into $2 billion is not treated the same as turning $1 billion into $2 billion. The score subtracts the cost of each investment, rewarding capital efficiency and penalizing strategies that spend heavily to produce a small number of headline wins. In the dataset, roughly three-quarters of investments produced negative net gains.
The third is value-add. Investors who lead rounds and take board seats receive extra credit relative to passive check writers.
The article also puts weight on how credit is split between firms and individuals. Investors move from one firm to another, so the model does not leave all the credit with the original platform when a partner later changes firms. One-quarter of the credit goes to the firm where the investment was made, while three-quarters goes to the partner’s current firm. Strebulaev says that reflects academic evidence suggesting most return dispersion comes from individuals rather than institutions.
Sequoia leads the 2026 ranking
Sequoia ranks first with 10,158 points. Andreessen Horowitz is second with 8,292. Accel, DST Global and Tiger Global make up the rest of the top five.
Strebulaev highlights two lesser-known names in the top 20 that founders building target lists may want to revisit. Parkway Venture Capital, founded in 2019, ranks No. 19 on the strength of Figure AI. Notable ranks No. 20. Neither is a household name, yet both place ahead of firms with much broader brand recognition.
The ranking table lists each firm’s position, headquarters, founding year, score and top deal. The top deal is the firm’s highest-scoring single investment under this methodology, which is not necessarily the best-known company in its portfolio or the one with the highest valuation.
Scores are rounded to whole numbers. Geographically, 62 of the top 100 firms are headquartered in California, 19 in New York, and six each in Massachusetts and Texas. Strebulaev writes that whatever expansion has occurred at the seed and angel level, the institutional core of U.S. venture capital remains in its historical centers.
The gap between top firms and the rest is huge
One of the clearest patterns in the ranking is the size of the drop-off. By No. 10, scores are already down to around 3,000, less than one-third of Sequoia’s level. By No. 100, the score falls to 245. The No. 1 firm’s score is about 41 times that of the No. 100 firm.
Strebulaev presents that as power-law behavior at the firm level, not just at the individual deal level. For founders, the practical point is that the difference between a top-decile investor and a merely well-known one can be an order of magnitude.
Unicorn counts say very little on their own
The ranking is not a unicorn-count contest. SV Angel backed about 139 unicorns and ranks No. 31. Insight Partners backed about 124 and ranks No. 28. Felicis has 58 unicorns and ranks No. 74. By contrast, DST Global has 62 unicorns and ranks No. 4, while Thrive has 47 and ranks No. 8.
The reason, according to the article, is that the model rewards actual value captured after dilution, not the number of billion-dollar logos on a firm’s website. A tiny early check in a company that later becomes a unicorn can still score poorly if it gets heavily diluted and ends up worth little at exit. A concentrated, board-level position in a smaller number of winners can generate far more points. On a score-per-unicorn basis, high-volume firms are around 6 points, while concentrated firms can exceed 80.
For founders, Strebulaev says that gap helps distinguish investors who commit and stay involved from those who write many small checks hoping one or two work out.
Older brands still matter, but newer firms can break in
Bessemer is the oldest name in the top 100, with venture roots tracing back to the 1970s. Inflection Ventures, founded in 2022, is the youngest. Eight of the top 20 were founded before 2000, evidence of how durable venture brands can be.
But longevity is not a requirement. Twenty-one firms in the top 100 were founded in 2015 or later, and some climbed quickly because of a recent investment that appreciated fast.
The ranking also includes specialist life sciences firms that made the list through therapy bets rather than software, including OrbiMed at No. 27, Atlas Venture at No. 38, ARCH at No. 49, Versant at No. 61 and Sofinnova at No. 75. Crypto-native firms also appear: Paradigm at No. 34, Pantera at No. 76, Multicoin at No. 84 and Polychain at No. 94. Strebulaev says the framework does not favor one sector over another; it measures value created net of cost and decay, wherever that value shows up.
That matters for founders in specialized categories. The best investor for a given startup may be a domain expert who never appears near the top of broad media lists.
AI is already reshaping the top of the list
Among the top 100 firms, 23 list a frontier AI or AI infrastructure company as their highest-scoring investment. That is nearly a quarter of the ranking. Many of those companies did not exist five years ago or were still small at the time, the article says.
Because the model includes a decay factor, the reshuffling happens in real time rather than years later. Strebulaev cites Thrive’s investment in OpenAI, Menlo’s investment in Anthropic and Lightspeed’s investment in Mistral as examples of early, concentrated bets that are rewarded quickly. In the article’s Q&A, he adds that OpenAI was named by four firms as their top investment, xAI by three, and Anthropic and Perplexity by two each.
What a ranking cannot tell a founder
Strebulaev warns against using the list as a simple spreadsheet for emailing firms from No. 1 to No. 100. Rankings are appealing because they appear to do the thinking for you, he writes, but the ordering is the least interesting part. What matters is the reasoning underneath and whether that reasoning matches what a founder actually needs.
He says three things should guide the final choice, and no score can fully capture any of them.
- Fit matters, not just rank. Cross-over and hedge-fund-style firms such as Tiger Global (5), DST Global (4), Dragoneer (23), Altimeter (24), Coatue (29) and Greenoaks (44) often buy large minority stakes and rarely take board seats. Sequoia and Benchmark follow a deep-engagement, board-heavy model. Sutter Hill (26) comes close to incubating companies from scratch. All of those approaches can score well, but they feel very different from a founder’s seat.
- Due diligence the partner, not the brand. One of the article’s strongest observations is that fully half of the top 100 firms have no partner in the top 100 individual investor ranking. A strong firm is not the same thing as a strong person. Founders should ask who will actually work with them, what else that person is handling and how long they have been at the firm, then speak with founders they backed, including founders of failed companies.
- Check behavior in down rounds. Strebulaev says how an investor behaves when a company is struggling is among the most important things a founder can learn, and no ranking can supply that answer.
He compares the founder-investor relationship to a marriage rather than a transaction. Investors can stay on the cap table or board for seven years, ten years and sometimes fifteen. A founder can change products, teams or even strategy, but removing an investor from the ownership structure or board is much harder.
The article also notes that value-add is included because deep involvement clearly affects outcomes. Still, the same board seat that opens one door can also block a sale, overturn a strategy or replace a chief executive. Founders should not only ask whether an investor can help them win, but also whether they want that person in the room on the company’s worst day.
How Strebulaev says founders should use the list
- Screen for performance first. If 5% of investors create about 90% of profits, they are worth real effort to reach.
- Filter for fit next. That includes stage, sector, check size and the kind of involvement the founder wants.
- Go one layer deeper and focus on the individual. A firm gets you the meeting; a single partner usually defines the relationship afterward.
- Run downside-reference checks. Speak with founders whose companies struggled, not only with the success stories featured on firm websites.
What comes next
Strebulaev says the same six-factor framework will be extended from firms to individual investors. In his 2026 top 100 individual venture ranking, more than half do not appear on that year’s Forbes Midas List at all.
He also says the project will expand internationally, publish sector rankings beginning with biotech and AI, incorporate verified data on both firms and investors, and release roughly 25 years of historical series so the rise and decline of firms can be seen directly.
Key figures restated in the article’s Q&A
- Sequoia leads the 2026 Strebulaev-Jackson ranking with 10,158 points, ahead of Andreessen Horowitz at 8,292 and Accel at 4,576.
- The No. 1 firm scores about 41 times the No. 100 firm, and by No. 10 the score is already below one-third of Sequoia’s.
- Backing more unicorns does not automatically mean stronger performance. SV Angel’s roughly 139 unicorns place it No. 31, while Thrive’s 47 put it at No. 8.
- About 5% of venture firms generated roughly 90% of the industry’s profits.
- Twenty-one of the top 100 were founded in 2015 or later.
- Of the top 100 firms, 62 are based in California, 19 in New York, and six each in Massachusetts and Texas.
For founders, the article’s bottom line is not to copy a ranking into a cold-email sheet and call it a strategy. Use the list to identify who has actually created value, then figure out whether that person or firm fits the company you are trying to build and the board relationship you are willing to live with for years.
