Five Names on the Door
On a Wednesday evening in September 2017, Bill Gurley walked out of an Uber board meeting and into the particular kind of silence that descends when a partnership has staked its reputation on a bet it cannot take back. Benchmark Capital had just filed a lawsuit against Travis Kalanick—the founder it had funded, the founder whose company it had nurtured from a black-car experiment into a $70 billion juggernaut, the founder it was now accusing of fraud. The suit alleged that Kalanick had deceived the board to expand his control, concealed ethical violations, and betrayed his fiduciary obligations to investors. Within hours, other venture firms were calling Benchmark's partners to express something between shock and horror. A coalition of Uber shareholders demanded that Benchmark relinquish its board seat and divest its shares. The letter was dressed in the language of governance reform but carried an unmistakable threat: this is what happens to VCs who turn on a founder.
"All we got was a shitload of flak from within the industry for doing it," one of Gurley's colleagues told The New Yorker. "Lesson learned. That's what you get for trying to do what's quote-unquote 'the right thing.'"
The irony was almost too neat. Benchmark had built its entire franchise on the proposition that a venture firm could be both deeply founder-aligned and uncompromising about governance—that these were complements, not contradictions. It was the firm that had turned a $6.7 million check into eBay into $5 billion in realized gains. The firm that backed Instagram, Twitter, Snapchat, Uber, Yelp, Zillow, Discord, and Elastic. The firm that had, alone among Silicon Valley's elite partnerships, refused to grow: five general partners, no growth fund, no hedge fund, no crypto fund, no media empire, no hundred-person platform team. The same $425 million fund size, vintage after vintage, while Andreessen Horowitz swelled past $35 billion in assets under management and Sequoia restructured itself into a permanent capital vehicle spanning public and private markets across three continents.
That refusal—to scale, to diversify, to become an institution—is the fact that explains everything else about Benchmark. It explains the returns (its 2011 fund reportedly grew investors' capital by roughly 25x). It explains the culture (equal economics among partners, with no founder or senior partner taking a larger share). It explains the specific texture of Benchmark's relationship with founders, which has at times been the most intimate and productive in venture capital and at other times the most combative. And it explains why a firm that manages less capital than most of its competitors' side vehicles remains, three decades after its founding, the entity against which every serious venture firm in the world measures itself.
The question Benchmark keeps being asked—will you ever grow?—is the wrong question. The right question is whether a model built on radical constraint can survive an industry that has made radical expansion the default setting.
By the Numbers
The Benchmark Machine
$425MConsistent fund size (2013, 2018, 2020 vintages)
~5General partners at any given time
~25xReported multiple on 2011 vintage fund
$5B+Realized gains from eBay alone
1995Year founded in Menlo Park
0Growth-stage or crossover funds raised
EqualCarried interest split among all partners
The Counter-Position
Benchmark was born from an argument about what venture capital had become. By the mid-1990s, the industry's two dominant franchises—Kleiner Perkins Caufield & Byers and Sequoia Capital—had achieved a kind of gravitational permanence on Sand Hill Road, but they'd also drifted into patterns that a handful of younger investors found self-defeating. Kleiner was riding
John Doerr's extraordinary instincts (he'd backed Amazon, Google, Sun Microsystems, Compaq) but operated with a hierarchical structure in which junior partners ate scraps while senior partners feasted. Sequoia under Don Valentine was brilliantly disciplined but similarly top-down. Both firms were beginning to layer on assets—larger funds, later-stage vehicles, international outposts—creating internal complexity that diluted focus.
Four men looked at this landscape and saw an opening. Bob Kagle had been at Technology Venture Investors. Bruce Dunlevie came from Merrill, Pickard, Anderson & Eyre. Andy Rachleff had co-founded a small firm but craved something more ambitious. Kevin Harvey was an entrepreneur who'd built a software company. They were joined shortly after by the person who, more than anyone, would become the firm's public face: Bill Gurley, a six-foot-nine former college basketball player from Dickinson, Texas, who had been a Wall Street research analyst at Deutsche Morgan Grenfell before the siren song of venture capital pulled him west.
What these five shared was not a sector thesis or a pattern-matching algorithm. What they shared was a structural conviction: that the venture capital firm itself was a product, and that the existing products were badly designed. The insight was that the most important variable in venture returns was not deal flow, not brand, not the ability to write large checks—it was the quality and intensity of the individual board member sitting across from the founder. And the way to maximize that quality was to strip everything else away.
We counter-positioned against one of the two major incumbents at that time, which was Kleiner Perkins.
The design choices flowed from this premise with almost mathematical precision. Five partners, not twelve or twenty. Equal carry—every partner received the same share of profits, regardless of seniority or which deals they'd personally sourced. Small funds—initially around $85 million for the first fund in 1995, rising only modestly over the decades—so that each partner could sit on a manageable number of boards and devote real time to each company. No associates, no principals, no junior investment professionals whose ambition might distort the decision-making of the partnership. And, critically, no "founding partner" hierarchy: no one got extra economics for having started the firm, which meant no one had structural leverage over anyone else.
This last detail was the most radical. In a world where Kleiner was synonymous with Doerr and Perkins, where Sequoia was Valentine and then Moritz, Benchmark was proposing an ego-free partnership in an industry that ran on ego. The name itself was a statement—"Benchmark" referred to no person, carried no founder's surname. It was a standard against which to measure, nothing more.
The eBay Proof
For a venture firm, the difference between an interesting philosophy and a legendary franchise is usually a single deal. For Benchmark, that deal was eBay.
In 1997, eBay was a small online auction site run by Pierre Omidyar, who had coded the original version as a side project while working at another company. The business was already profitable—almost absurdly so, given the internet-bubble climate in which profitability was considered an irrelevant distraction. Bob Kagle led the deal for Benchmark. The firm invested approximately $6.7 million for a 21.5% stake in what was then a company with a handful of employees operating out of modest offices in San Jose.
What happened next became the foundational myth of the Benchmark model. The partners didn't just write a check and show up to quarterly board meetings. Bruce Dunlevie, who joined the board, was instrumental in recruiting
Meg Whitman as CEO—an executive hire that transformed eBay from a quirky marketplace into a global platform. Benchmark partners helped professionalize every dimension of the operation: finance, marketing, operations, the legal framework for what was an entirely novel transactional model. When eBay went public on September 21, 1998, its shares opened at $53.50, more than double the offering price. Within months, Benchmark's $6.7 million stake was worth over $5 billion.
The magnitude was staggering, but the mechanism was what mattered to the partnership's self-conception. This wasn't a spray-and-pray bet that happened to hit. It was a concentrated investment in a company where Benchmark partners had been deeply involved in operational decisions, had recruited the management team, and had shaped the strategic direction. The returns validated the structural thesis: small fund, concentrated bets, deep engagement.
Benchmark's foundational investment
1997Benchmark invests ~$6.7M in eBay for a 21.5% stake.
1998eBay IPOs on September 21. Benchmark's stake exceeds $5 billion in paper value.
1999Benchmark distributes 13.8 million eBay shares to its limited partners via a pro rata distribution on December 7.
On December 7, 1999, Benchmark Capital Partners, L.P. and Benchmark Founders' Fund, L.P. executed a pro rata distribution to their limited partners of 13,775,508 shares of eBay common stock. The SEC filing documenting this distribution—a dry, tabular document listing hundreds of names and share counts—is, in its own way, one of the most remarkable documents in venture capital history. It is the ledger of a single deal returning more than fifty times the entire fund to its investors.
The eBay windfall created a problem that would define Benchmark's next two decades: how do you follow an act like that without succumbing to the temptation to change the formula?
The [Discipline](/mental-models/discipline) of Smallness
The late 1990s and early 2000s destroyed most of the lessons the venture industry thought it had learned. The dot-com bubble inflated, popped, and left behind a landscape of wrecked firms and chastened investors. Benchmark was not immune to the carnage—it had made bets that went to zero, including a painful foray into Webvan, the grocery delivery company that became a punchline for bubble-era hubris. But the partnership emerged from the wreckage structurally intact, in part because the eBay returns provided an enormous cushion and in part because the small-fund model limited exposure to any single disaster.
What happened next was more telling. In the late 1990s, flush with eBay's billions, Benchmark had briefly expanded—opening offices in London and Israel, launching Benchmark Europe with a separate team. It was the obvious move: the internet was globalizing, and international markets represented untapped opportunity. Almost every top-tier venture firm was doing the same.
It didn't work. The international offices created coordination problems, diluted the partners' attention, and—most importantly—violated the core structural principle that had generated Benchmark's returns in the first place. By the early 2000s, Benchmark had shuttered its international operations. The lesson was seared into the partnership's institutional memory: expansion destroys the thing that makes this work.
This became the template. Every few years, as the venture industry underwent another cycle of expansion—first the Web 2.0 era, then the growth-equity boom, then the mega-fund era kicked off by Andreessen Horowitz, then the SoftBank Vision Fund's hundred-billion-dollar experiment, then the 2021 ZIRP-fueled frenzy—someone would ask Benchmark the same question. Will you raise a growth fund? Will you add more partners? Will you launch a crypto vehicle?
The answer was always no.
At those stages, capital is almost certainly not the limiter. It's certainly attractive, you know, larger fund size and to write larger checks. It is alluring, but it's actually not what you need at that point in time.
The consistency is almost suspicious. Benchmark raised $425 million for its 2013 vintage fund. It raised $425 million for its 2018 vintage fund. It raised $425 million for its 2020 vintage fund. Adjusted for inflation, the fund size has actually shrunk in real terms over the past decade—at a time when Andreessen Horowitz was raising $9 billion mega-funds and Tiger Global was deploying billions per quarter into late-stage private companies.
The math behind this discipline is instructive. A $425 million fund with five partners means each partner is responsible for deploying roughly $85 million across a portfolio of perhaps six to ten companies over three to four years. That is an extremely concentrated portfolio by venture standards. It means every single investment decision carries enormous weight—and every single board seat demands genuine engagement. There is no "portfolio theory" diversification available at this scale. Benchmark cannot afford to make thirty bets and hope three pay off. Every company must be treated as if it could be the one that returns the fund.
This is the structural logic that produces both Benchmark's extraordinary returns and its extraordinary willingness to intervene in its companies. When your fund depends on a handful of outcomes, you cannot afford to be a passive, "founder-friendly" capital provider who looks the other way when things go wrong.
The Tall Men and Their Successors
Randall Stross's
eBoys, published in 2000, gave the world its first interior portrait of Benchmark—and inadvertently supplied the firm with a mythology it neither sought nor could quite escape. The "six tall men" of the subtitle (the fifth original partner, Kevin Harvey, was later joined by others) were depicted as a band of equals, collegial and competitive in roughly equal measure, operating with an intensity and mutual accountability that was unusual even by Silicon Valley standards.
The book captured a moment. But Benchmark's most important institutional innovation—the one its competitors found hardest to replicate—was its approach to succession. Unlike Kleiner Perkins, which spent two decades struggling to groom successors for Doerr and eventually fractured (Mary Meeker's growth fund outperforming the venture team created internal resentments that accelerated the firm's decline), Benchmark built succession into the model from the start.
The equal-carry structure was the key. Because no founding partner received economics that made them irreplaceable, the firm could bring in new partners without creating the kind of two-tier system that breeds resentment. When a partner decided to step back—as Andy Rachleff eventually did, leaving to found Wealthfront, or as Bob Kagle did—the partnership could elevate or recruit a replacement who would immediately receive the same economics as every other partner. No ramp-up period. No junior carry. Full equality from day one.
The succession track record is remarkable for its consistency. Peter Fenton, who joined Benchmark from Accel Partners, became one of the most successful venture investors of the 2000s and 2010s, leading investments in Twitter, Yelp, and Elastic. When Fenton joined, he received the same carry as the founders. When he recruited Sarah Tavel—who had been a product manager at Pinterest before moving into venture—she received the same carry as Fenton. When Miles Grimshaw came over from Thrive Capital in 2020, at age thirty, the same.
The pattern was not accidental. Benchmark's partners explicitly modeled their succession framework on professional partnerships like Goldman Sachs in its pre-public era, where partnership was a genuinely shared enterprise rather than a fiefdom.
Key Benchmark general partners across eras
1995Founding partners: Bob Kagle, Bruce Dunlevie, Andy Rachleff, Kevin Harvey. Bill Gurley joins shortly after.
2006Peter Fenton joins from Accel Partners, receiving equal carry.
2009Peter Fenton leads Benchmark's investment in Twitter.
2013Matt Cohler, formerly VP of Product at Facebook and LinkedIn, is a general partner during this era.
2015Mitch Lasky, former gaming executive (Activision, EA, Jamdat), brings sector expertise.
2017Sarah Tavel joins as general partner after stints at Greylock and Pinterest.
2019Chetan Puttagunta joins, bringing deep enterprise and open-source expertise.
The incoming partners share a pattern: operational experience before venture, intellectual seriousness bordering on obsessiveness, and a willingness to subordinate personal brand to institutional identity. Mitch Lasky had been an executive at Electronic Arts, Activision, and Jamdat Mobile before joining Benchmark. Chetan Puttagunta brought deep expertise in enterprise software and open-source business models. Eric Vishria had been COO of Optimizely. These are not career VCs who climbed the associate-to-partner ladder; they are people who built things before they funded things.
The outgoing partners, crucially, actually leave. They don't linger as "venture partners" or "senior advisors" or "chairman emeritus" collecting a piece of the carry while occupying space in the partnership. When Andy Rachleff left, he left. When Bill Gurley stepped back from day-to-day investing, he did so without clinging to an economic tail that would dilute the next generation. This is, in venture capital, almost unheard of. The industry is littered with firms whose founding partners hung on too long, creating a gerontocracy that blocked talented younger investors from ascending—or, worse, that maintained a two-tier economic structure in which the founders took the lion's share of profits while the people doing the actual work took the scraps.
The Persuasion Business
Bruce Dunlevie once described himself as being in the "persuasion business." It's a more revealing phrase than it might first appear. Venture capital, at its most effective, is not primarily a financial activity—it is an act of sustained influence exercised across years of board meetings, crisis phone calls, recruiting conversations, and strategic debates with founders who are simultaneously brilliant and unreasonable.
Dunlevie embodied this role with unusual intensity. A six-foot-four former high school quarterback from Texas with a history degree from Cambridge, he was the Benchmark partner most likely to quote Lord Acton at a board meeting and the one most likely to threaten to break a founder's arm if the situation warranted it. Both of these things happened at WeWork.
The WeWork saga, chronicled with devastating precision in Reeves Wiedeman's
Billion Dollar Loser and in Charles Duhigg's reporting for
The New Yorker, is the case study that haunts Benchmark—not because the firm lost money (it didn't; Benchmark reportedly earned a return of more than 1,000% on its WeWork investment), but because it exposed the limits of the "persuasion business" model when applied to a founder operating without constraint.
Benchmark led WeWork's $17 million Series A in 2012, at a $97 million post-money valuation. Dunlevie joined the board and became a mentor to Adam Neumann. At that early stage, the dynamic worked as designed: an experienced, deeply engaged board member guiding a charismatic but undisciplined founder. Dunlevie admitted to a partner that he wasn't certain how WeWork would ever become profitable, but he was taken with Neumann. The calculation was characteristic of venture capital's fundamental bet: back the person, trust that the person will figure out the business model.
The trouble began when Neumann's charisma proved so effective at raising capital that it became impossible to constrain him. By 2014, Neumann was demanding majority voting control over the board—a governance structure that would make him virtually unaccountable. Dunlevie objected. He quoted Lord Acton: "Power tends to corrupt, and absolute power corrupts absolutely." Neumann said he didn't care about Lord Acton. No other board member supported Dunlevie's objection. And here was the moment where Benchmark's structural constraint—small fund, concentrated bet, enormous exposure to a single company's outcome—created a perverse incentive. Walking away from a company whose valuation had gone from $97 million to $15 billion would have been, as one colleague put it, "the stupidest fucking idea on earth."
Dunlevie stayed. The board gave Neumann his voting control. Over the next five years, Benchmark watched as WeWork's internal culture deteriorated into something between a fraternity and a cult, as Neumann spent $60 million on a corporate jet and $13 million of company funds on artificial-wave pools, as the company's losses reached $219,000 per hour, and as the S-1 filed ahead of the planned 2019 IPO opened with a dedication "to the energy of We" and closed with photographs of a tropical forest that Neumann and his wife had promised to protect.
Our job had basically become to make sure Adam didn't do anything really stupid or really illegal—the board knew Adam was the key to raising money, and, as long as their valuations kept going up, they weren't going to risk upsetting him.
When the S-1's public release in August 2019 triggered universal derision, and when the Wall Street Journal reported that Neumann had smoked marijuana on a private jet and told people he planned to become "president of the world," the board finally acted. Dunlevie delivered the ultimatum at a dinner. He told Neumann he was toxic, threatened to break his arm if he didn't resign, and warned that his loans backed by WeWork stock would be called in, ruining him. Neumann stepped down on September 24, 2019.
The aftermath was instructive. Benchmark walked away with roughly $300 million—a return of more than ten times on its investment. Neumann received about $725 million. Over ninety percent of WeWork's current and former employees saw their stock and options rendered effectively worthless. The board had approved every one of Neumann's proposals unanimously, had never formally dissented in the meeting minutes, and had, in the words of one executive, operated as "a Vichy board."
What makes this story central to understanding Benchmark—rather than merely embarrassing—is that the same structural features that produced the WeWork outcome also produced the eBay outcome. Concentrated bets. Deep board engagement. Willingness to get close to founders. The model's strength is inseparable from its vulnerability. When the founder is Pierre Omidyar or
Evan Spiegel, intimacy produces extraordinary alignment. When the founder is Adam Neumann, it produces complicity.
Suing the Founder
The Uber lawsuit was the other side of the coin—and in many ways the more important story for understanding Benchmark's identity.
Bill Gurley led Benchmark's 2011 investment in Uber, joining the board when the company was still a scrappy black-car service in San Francisco. Over the next six years, Gurley became one of Uber's most consequential directors, helping shape strategy, recruiting executives, and providing the kind of operational counsel that Benchmark's model was designed to deliver. But he also watched as Travis Kalanick—brilliant, relentless, and constitutionally incapable of self-regulation—built a culture of rule-breaking that extended far beyond competitive aggression into territory that was, by any measure, toxic.
The litany is familiar: systematic sexual harassment documented by Susan Fowler's viral blog post, surveillance programs targeting regulators and competitors, a culture that celebrated intimidation, the concealment of a massive data breach, executive misconduct at virtually every level. Gurley and Benchmark publicly supported Kalanick for years as these issues festered—because the company was winning, because the valuation was rising, and because the "founder-friendly" norm made any public criticism a professional death sentence.
When Benchmark finally broke—joining with other investors including First Round Capital, Lowercase Capital, Menlo Ventures, and Fidelity to push Kalanick out in June 2017—it was too late for the move to be read as principled governance. And when Benchmark followed up with the lawsuit in August, alleging fraud and breach of fiduciary duty, the backlash was immediate and severe. Other shareholders demanded that Benchmark give up its board seat. The venture community largely sided with the sentiment, if not the specifics, of the criticism: you don't sue your founders.
Benchmark wrote a nearly 1,000-word letter to Uber employees. "Perhaps the better question is why we didn't act sooner," the letter read. It was a remarkable document—frank, almost anguished, and openly grappling with the tension between loyalty to a founder and obligation to a company. "We believed then, as we believe now, that failing to act would have meant endorsing behavior that was utterly unacceptable in any company, let alone a company of Uber's size and importance."
Kalanick's response was swift: "Like many shareholders, I am disappointed and baffled by Benchmark's hostile actions, which clearly are not in the best interests of Uber and its employees on whose behalf they claim to be acting."
The lawsuit was eventually dropped, Kalanick left the board, and Uber went public in May 2019 at a valuation of approximately $82 billion. Benchmark's stake was worth billions. But within the partnership, the experience left a complicated residue. The firm's willingness to act was, depending on your perspective, either the most courageous governance intervention in modern venture capital history or a cautionary tale about what happens when a VC firm prioritizes its own financial interests and tries to dress it up as principle.
The truth is probably both, which is exactly the kind of ambiguity that Benchmark's model produces. A firm with five partners and a concentrated portfolio cannot afford to be passive when a company goes off the rails. But it also cannot afford to be seen as founder-hostile in an industry where deal flow depends on reputation. The Uber lawsuit made Benchmark the only elite venture firm in Silicon Valley that had actually sued a founder. Some admired the nerve. Many more filed it away as a reason to think twice before taking Benchmark's money.
The Twitter Bet and the Art of Quiet Conviction
Not every Benchmark story is a governance drama. Many of the firm's most important investments followed a simpler, more elegant arc: identify a product with extraordinary organic adoption, invest early, provide genuine board-level support, and hold through the volatility.
Twitter was the paradigm case. Peter Fenton, who had been watching the company for two years, led Benchmark's $35 million investment in February 2009, joining the board alongside Todd Chaffee of Institutional Venture Partners. At the time, Twitter had 29 employees, had never earned a single dollar of revenue, and was growing at a rate that confounded even its own founders. Active users had increased 900% in the previous year.
"They decided to partner with an active West Coast syndicate to vigorously pursue the path of independence," Fenton said at the time. "As a business opportunity, it jumped out to us as having many potential revenue streams that support, and don't undermine, its success." The statement was characteristically measured for a Benchmark partner—analytical rather than hyperbolic, focused on business model optionality rather than world-changing rhetoric.
When Twitter filed its S-1 in 2013, the extent of Benchmark's conviction became visible. Fenton held 31.5 million shares on behalf of Benchmark Capital. At the company's August 2013 fair-value estimate of $20.62 per share, that stake was worth approximately $650 million—before the IPO pop that would push it considerably higher. Twitter went public on November 7, 2013, opening at $45.10 per share.
The Twitter investment also revealed something about Benchmark's sourcing model that is easy to overlook: the firm does not operate a massive scouting network or employ armies of associates combing through Y Combinator batches. Its sourcing depends on five partners who are deeply embedded in the technology ecosystem, each maintaining a dense network of founder relationships, and each making a small number of high-conviction bets per year. Fenton found Twitter not through a proprietary algorithm but through two years of patient observation—watching the product, talking to users, tracking adoption curves—before the company was even looking for West Coast venture capital.
This is the opposite of the SoftBank model, in which Masayoshi Son could commit $4.4 billion to WeWork after a twelve-minute tour. It is investment as craft rather than investment as leverage.
The Snapchat Conviction and the $3 Billion No
Benchmark's investment in Snapchat followed a similar pattern of early conviction in a product that most of the investment establishment dismissed. Mitch Lasky, the former gaming executive who joined Benchmark, led the firm's Series A investment in Snap in 2013. The bet was characteristic Benchmark: a consumer product with explosive organic growth, a founder (Evan Spiegel) with extraordinary product instincts, and a business model that didn't yet exist.
What made the Snapchat investment distinctive was what Spiegel did in November 2013, shortly after Benchmark's investment: he turned down a $3 billion acquisition offer from Facebook. Spiegel was 23 years old. He was walking away from a guaranteed fortune on the conviction that Snapchat could become something bigger. It was exactly the kind of founder audacity that Benchmark's model was designed to support—and exactly the kind of decision that would have terrified a VC firm with a larger fund and less tolerance for concentrated risk.
Snap filed its S-1 in February 2017, disclosing its intention to go public at a valuation of approximately $25 billion. By that point, the company had over 150 million daily active users and had reinvented itself as a "camera company" with the launch of Spectacles. Benchmark's early stake, purchased at a valuation many orders of magnitude lower, had become one of the firm's signature returns.
The Snapchat investment also highlighted a tension within Benchmark's model. The firm's equal-carry structure meant that Lasky's success with Snap benefited every partner equally—which created powerful alignment but also meant that a single partner's mediocre fund-period performance was subsidized by a colleague's genius bet. The structure demands that every partner maintain an exceptionally high standard, because there is no carrying passengers. If one partner's portfolio drags, the economics are shared. This creates intense peer accountability—a kind of internal tournament that substitutes for the hierarchical oversight found at more traditional firms.
The Anti-Platform
In 2009,
Marc Andreessen and
Ben Horowitz founded Andreessen Horowitz—a firm that was, in many ways, the anti-Benchmark. Where Benchmark had five partners and no support staff, a16z launched with an explicit "platform" model: dozens of operating partners, recruiters, marketing specialists, and communications professionals who would provide portfolio companies with services that went far beyond a board seat and a phone call. Where Benchmark raised small funds and stayed in the early-stage trenches, a16z raised progressively larger vehicles and expanded into growth, bio, crypto, and infrastructure. Where Benchmark's partners kept relatively low public profiles (Gurley's blog and occasional conference appearances being the notable exception), Andreessen became arguably the most prolific content producer in the history of venture capital—podcasts, blog posts, manifestos, tweet storms.
The two firms were waging an ideological war over the nature of venture capital itself. Andreessen explicitly framed a16z as a correction to what he saw as the complacency of the old guard. "What we did at Benchmark about ten years ago when we started Benchmark," Horowitz said on Acquired, referring to the counter-positioning move—but the implication was clear: a16z was now counter-positioning against Benchmark's asceticism.
What we did at Benchmark about ten years ago when we started Benchmark...we counter-positioned against one of the two major incumbents at that time, which was Kleiner Perkins.
The competitive dynamics were real. a16z's platform model was designed, in part, to win deals that Benchmark would otherwise have won by offering founders a more comprehensive package of support. If Benchmark's pitch was "you get a world-class board member who will be in the trenches with you," a16z's pitch was "you get a world-class board member plus a hundred-person team that will help you recruit, do PR, lobby regulators, and build your go-to-market engine."
For Benchmark, this posed an existential question: could five partners with no support infrastructure continue to win the best deals against a firm that offered founders an entire institution? The answer, empirically, has been yes—but the margin has thinned. Benchmark continues to win its share of the most competitive Series A deals, but the firm's advantage now depends almost entirely on the individual reputations of its partners and the track record of its brand. A founder choosing between Benchmark and a16z is making a choice about what kind of help they want: the intensive mentorship of a single deeply experienced board member, or the comprehensive support of a platform that can solve fifteen different operational problems simultaneously.
The honest answer is that different founders need different things at different stages, and that Benchmark's model has natural limits when a company reaches the scale at which a single board member—however brilliant—cannot provide all the help that's needed. Benchmark's portfolio companies inevitably hire other firms' partners to their boards in later rounds. The question is whether that early-stage relationship, forged in the intimacy of a first institutional round, creates enough loyalty and enough strategic advantage to justify the model's constraints.
The Dinner and the Doctrine
Benchmark's internal culture is organized around a ritual that encapsulates the entire philosophy: the Monday partner meeting. Five people in a room, no associates to present memos, no investment committees to ratify decisions. Every partner sees every deal. Every partner has an equal vote. A single enthusiastic partner can champion an investment, but any partner can also kill one.
The meeting is breakfast, then lunch, then often dinner—a single continuous conversation that can stretch for hours. The partners review existing portfolio companies, debate new investment opportunities, and, crucially, argue. The equal-carry structure means that disagreement carries no career risk; you cannot be punished for dissenting, because there is no senior partner to punish you. This produces, by all accounts, a quality of debate that is unusual in venture firms, where junior partners learn quickly that contradicting the founder is unwise.
The meeting also serves as the mechanism for the most consequential decision a venture firm makes: when to say no. Benchmark's small fund size and concentrated portfolio mean that the opportunity cost of every investment is enormous. Saying yes to one company means saying no to three others. The partnership's ability to maintain discipline in the face of FOMO—the fear of missing out that drives most venture capital decision-making—is arguably more important than its ability to pick winners.
This discipline has produced notable misses. Benchmark did not invest in Facebook, Airbnb, or Stripe—three of the most valuable venture-backed companies of the past two decades. These misses are the price of concentration. A firm that makes ten investments per fund simply cannot cover the entire surface area of innovation. What it can do is generate enormous returns from the investments it does make, which is why Benchmark's fund multiples have remained among the highest in the industry despite missing individual companies that returned more than its entire fund.
What the Small Fund Teaches
Sebastian Mallaby's
The Power Law devotes significant attention to the structural dynamics of venture capital, and Benchmark features prominently as the exception that proves the rule: in an industry that has been consolidating around ever-larger pools of capital, the firm has demonstrated that discipline and constraint can generate returns that dwarf those of firms managing ten or twenty times more money.
The power-law distribution that governs venture returns—where a small number of investments generate the vast majority of profits—should, in theory, favor larger funds that can make more bets. If returns follow a power law, buying more lottery tickets should increase your expected value. This is the logic behind SoftBank's Vision Fund, behind Tiger Global's indiscriminate deployment during 2021, behind every mega-fund that has been raised in the past decade.
Benchmark's counter-argument is that the power law operates at the level of the individual investment, not the fund. A $425 million fund that puts $30 million into a single company that returns 100x generates the same absolute dollar return as a $4.25 billion fund that spreads money across forty companies, one of which returns 100x. But the smaller fund achieves a dramatically higher multiple on invested capital—and because Benchmark's partners receive equal carry on a smaller base, the per-partner economics are extraordinary.
The constraint also produces a selection effect. When you can only make ten investments per fund, you are forced to be ruthlessly selective. You cannot "take a flier" on a speculative bet, because every flier comes at the cost of a conviction investment. This discipline, compounded over decades, produces a portfolio that is systematically skewed toward the highest-conviction opportunities.
Of course, this argument only works if the partners' conviction is well-calibrated—if they are, in fact, better at picking winners than the median venture investor. Benchmark's track record suggests that they are. But the firm's success is inseparable from the specific people who have occupied its five seats over the past three decades. The model is a machine for amplifying human judgment. When the judgment is good, the amplification is spectacular. When the judgment is flawed—as it was, arguably, in the WeWork boardroom—the amplification cuts the other way.
The Firm That Refused to Become a Company
Every venture firm eventually faces the question of what it wants to be when it grows up. Kleiner Perkins tried to become a multi-strategy asset manager and nearly destroyed itself in the process, hemorrhaging talent (Mary Meeker departed to launch her own fund, Bond Capital) and missing a decade of top-tier deals. Sequoia restructured itself into a permanent capital vehicle—the Sequoia Fund—before abandoning that structure in 2023 and splitting into three separate regional entities. Andreessen Horowitz hired hundreds of employees, launched podcasts and media properties, registered as a financial advisor, and began managing money in public markets.
Benchmark did none of these things. The firm in 2024 looks structurally identical to the firm in 1995: five general partners, a small fund, equal carry, early-stage focus. The office is modest. The website is spare. There is no podcast, no blog (Gurley's personal blog notwithstanding), no media empire. The partners do not publish annual letters or op-eds or manifestos about the future of computing.
This is not, as it might appear, an absence of strategy. It is the strategy. Benchmark's implicit thesis is that venture capital is an artisanal activity—closer to surgery or trial law than to asset management—and that the attempt to industrialize it destroys the thing that makes it valuable. The value of a Benchmark board seat comes not from the brand or the platform but from the specific human being sitting in the chair: their pattern recognition, their network, their willingness to make uncomfortable phone calls, their capacity to sit with a founder through a crisis.
This thesis has obvious limits. It depends on the continued ability to recruit extraordinary investors into a partnership that offers no empire-building opportunities, no chance to manage billions, no path to becoming an institution. The pitch to a potential Benchmark partner is: you will be one of five equals, you will manage a small fund, you will sit on a handful of boards, and you will do this for a decade or two before handing the seat to someone else. In an industry where the most ambitious investors are launching their own multi-billion-dollar firms, this pitch requires a particular kind of person—someone who values craft over scale, impact over empire, and the satisfaction of a single great investment over the thrill of managing a giant portfolio.
Benchmark has found these people, consistently, for three decades. Whether it can continue to do so—as the incentive structures of the industry increasingly favor scale—is the open question that no amount of fund performance can definitively answer.
The partnership keeps its fund at $425 million. Five names on the door. No associates. Equal carry. The industry keeps asking when this will change. And somewhere in Menlo Park, or maybe at a dinner that has lasted well into the evening, five people keep saying the same thing they have always said.
Nothing.
