Five People Managing $500M, Six Hundred Managing Billions: Where Does VC Money Actually Go

Deep thoughts on AI and aspirations —— ByteDance Deep Thought Circle

In 2009, right after Ben Horowitz closed his startup’s first round, investor David Beirne asked in front of his entire core team: When are you planning to hire a real CEO?

Ten years later, the man publicly questioned as “unfit to be CEO” co-founded Andreessen Horowitz (a16z) with Marc Andreessen, becoming Benchmark’s fiercest rival. The sting of Beirne’s comment had long since faded—Bill Campbell later called him the best CEO he ever worked with, and Campbell’s students included Jobs, Bezos, and Zuckerberg.

The feud between these two funds is just the opening act. What’s truly worth unpacking are the two completely opposite business models they represent: one firm uses five or six partners plus under $500 million in fund size to deliver some of the best returns of the past decade; the other uses an organization of hundreds and billions in management fees to become Silicon Valley’s startup services gateway. Both are hugely successful, so what did each pay to get there?

Let’s Start with the Numbers

Beyond the numbers are two telling details. Benchmark’s website has essentially one link—to their portfolio’s Twitter page. A16z brought in Michael Ovitz, founder of Hollywood talent agency CAA, as an advisor to come in one day a week and teach them how to run a network.

Benchmark partner Peter Fenton put it this way: “Compared to a marching band, we’re more like a jazz band. When you limit your scale, you have nowhere to hide.” This line is worth copying for anyone running a small team, but before you do, think carefully about what “nowhere to hide” actually means.

Small Isn’t a Virtue, Small Is a Pressure Cooker

The internet is full of hymns praising small team agility and big organization rigidity. Benchmark’s example gets cited constantly, as if being small is itself a moat.

This is a misreading. Being small doesn’t generate returns. Being small just transmits the pressure of having nowhere to hide directly to every decision.

Look at how they actually work. The day after Domo founder Josh James reached out, he faced Benchmark’s entire partnership team. They got straight to the point: We don’t want to invest $10 million, that’s nonsense; we won’t give you $20 million either, that’s just guessing at valuation; you’re Josh James, you need $30 million, and we want to give it to you right now. Terms printed within 30 minutes, money wired in 17 days. Note the mechanism here: five people must jointly commit on the spot, no junior partners to buffer due diligence, no committee process. If the judgment is wrong, all five are wrong together, with no structure to dilute responsibility.

The flip side of this equation is risk. Matt Cohler and Bill Gurley snagged the Uber investment by noticing Kalanick’s car service stopped at a competitor’s building—Cohler simply called the car away, making the competitor run over. This approach requires partners to be analysts, hunters, and the network itself simultaneously. Small organizations compress three roles into one person. Teams without sufficient talent density trying this model die faster than big organizations. When Fenton says “nowhere to hide,” he omits the second half: nowhere to hide means the partners’ true capabilities are fully exposed.

Scale has a second cost, which Benchmark experienced with both Google and Airbnb. Partner bandwidth is limited, and missed deals don’t get filled. Another line circulating in Silicon Valley circles: Benchmark still beats a16z on investment returns, but Benchmark only has a five-round magazine—when it’s empty, it’s empty.

a16z’s Calculation: Turning Service into Product

Now look at a16z’s choice from the other direction. It takes the management fee pool—traditionally split among partners—and spends heavily on post-investment services: dozens of people dedicated to helping portfolio companies find talent, find customers, do marketing, with a network claiming to cover a thousand executives and five thousand engineers. In Q1 2012 alone, they organized 88 enterprise customer briefings and 1,625 introductions.

Marc Andreessen’s statistical basis: 96% of annual returns in the venture capital industry come from the top 15 most successful investments. To claim a spot in those 15, money alone isn’t enough—you need to be the firm founders won’t sign with anyone else before meeting. Services buy you gateway position, gateway position buys you selection rights over the highest-quality deals.

This path has failed before. Mohr Davidow and Charles River both tried dedicated service teams with mixed results. The industry’s harsher version: when you go to a clinic, do you want to see the doctor or his assistant? A16z bet on the opposite judgment: what early-stage companies lack most isn’t judgment, it’s hiring and first customers—precisely what professional teams can deliver at scale.

The industry landscape a decade later provides a provisional answer. A16z became an unavoidable force in Silicon Valley, but two new problems emerged. After fund scale expanded, tens of billions couldn’t all go into early stage—a16z was forced into later stages, growth stages, multiple product lines. The organization became a company, and post-investment service’s marginal value diluted as portfolio company count grew. Meanwhile, the market spawned a wave of solo GPs and boutique funds. LP money flowed back to small, sharp models—a species that didn’t exist in 2012. The problem a16z solved (institutionalized service) and the problem it created (institutionalization itself) are two sides of the same coin.

Worth mentioning a judgment the source material didn’t expand on: a16z’s most enduring legacy may not be return rates, but how it expanded VC industry competition from judgment to organizational capability. Every new fund since has had to answer: what organizational model will you compete with? That question is a16z’s invention.

The Real Answer Is Hidden in Power Laws

Zoom out and you’ll find a counterintuitive fact: these two approaches seem opposed but actually share the same profit engine—venture capital’s power law structure. Andreessen himself calculated that 96% of returns concentrate in 15 deals. Benchmark’s decade of returns shows the same structure: 8 funds returned a combined $22.6 billion, 10x net of fees over ten years, driven by a few hits like Twitter, Snapchat, Uber, while most other projects contributed close to zero to total returns.

This fact flattens all the surface-level philosophical differences between the two firms. Both accept that 95% of portfolio projects are waste—it’s a cost both groups accept. The disagreement is only about how to increase the probability of hitting those 15 deals: Benchmark uses partner density to buy judgment quality, five people betting all capital on their own brains; a16z uses scale to buy coverage and gateway position, betting the hottest projects can’t bypass it.

Neither path transcends the power law itself. This is a reminder for all organizations learning “small team philosophy” or “platform approaches”: form is a product of incentive structure. Business model comes first, organizational form second. Benchmark can stay at five people because their profits come entirely from a few hits—management fees split equally among five top-tier partners is the optimal allocation. A16z can afford six hundred people because their strategy requires continuously locking down gateways to all hot deals—services are the cost of acquiring that lock. Learning either firm’s form without learning its business model is like buying the shell without the engine.

At this point, we can make a judgment. These two funds provide opposite sides of the same insight: organizations aren’t better smaller or better larger—organizations must align with their profit engine. If you eat by a few high-quality judgments, you should compress costs to the minimum and concentrate responsibility to the maximum. If you eat by coverage and gateway position, you should spend real money on scale. The most dangerous position is the middle—neither density nor scale, yet still collecting management fees. Most mediocre firms on Sand Hill Road stand exactly in this position.

Key points: Benchmark’s small isn’t a virtue but a pressure cooker—five people decide jointly with no dilution of responsibility; a16z’s service team is fixed cost buying gateway position; two opposite approaches share the same power law engine, 96% of returns concentrate in 15 deals; organizational form must align with profit engine, the most dangerous position is stuck in the middle.