How to scale a venture fund, lessons from FTX Ventures

Investing in Anthropic, Cursor, Figma, Circle, Solana, Sui by breaking the rules

Venture investing doesn’t scale. For an industry obsessed with investing in tech companies that scale through technology, venture is a tight network of individuals deploying relatively small amounts of capital (total VC AUM is ~$1T while private equity sits around $6T; global equities are around $78T). When venture firms add more partners or capital, they almost always perform worse in aggregate.

There’s a massive incentive for people to figure out how to scale venture firms. If a firm cracks this code can deploy capital across a wider array of companies and secure larger allocation into the winners. Numerous funds and platforms have tried different strategies to scale under their respective theses with varying degrees of success:

• AngelList: anyone can sign up and spin up a venture fund at cost
• Pioneer: the internet provides a unique substrate to identify talent globally via gamification and program metrics
• SoftBank: founders optimize for the best financial terms and offering better prices can generate outsized winners
• Y Combinator: young technical talent is underserved, business acumen is learnable on the job

Unlike many of their financial industry counterparts (investment bankers, consultants, etc.), venture capitalists are not known for their work ethic. Over the past decade, few firms have attempted meaningful innovations to scale venture funds. This can be partly attributed to the low personnel turnover rate and many less-than-ambitious people. Some are playing the management game, others spend twenty years coasting towards retirement, and some just aren’t that sharp.

FTX entering the venture game was a rare occurrence of an ambitious, resourced team deploying capital. FTX Ventures deployed over $5.2B in capital across more than 400 companies, with significant equity stakes in Anthropic, Cursor, Solana, and Sui. Today, that portfolio would be worth well over $20B.

FTX’s actions were illegal and inexcusable. But unlike most other venture firms, they employed a strategic, differentiated approach that leveraged their brand to deploy capital at scale. While they adopted some tactics from other firms, their implementation was a distinct point on the capital-strategy frontier and continues to influence venture firms today.

1. Get onto the field and operate, double down on your best theses and comparative advantages. The best returns to investment are your social capital.

FTX started with Alameda when they were trading on Binance market making and running a hedge fund. Instead of using money and social capital from Alameda to start a venture firm, they started FTX to compete with Binance as they saw the benefits of creating a centralized exchange. (Perhaps too many benefits.)

Matt Huang, general partner at Paradigm and investor in FTX, is now running Tempo. Huang is uniquely positioned to lead Tempo given his experience as the founder of one of the largest crypto venture firms along with the backing of Stripe (where he also sits on the board).

To be sure, FTX did not start this trend. Keith Rabois started OpenStore while working as a partner at Founders Fund. Fellow PayPal mafia member Max Levchin started Affirm while running his venture firm SciFi. FTX pushed this approach further, running multiple operations in parallel to create compounding value and showing that running multiple operations is a strategic advantage, not a distraction.

2. Invest in competitive markets and don’t be afraid to buy current beta.

In the vast majority of industries, the best-performing and most important businesses are obvious. In AI, these are the major research labs (OpenAI, Anthropic, DeepMind). In crypto, these are the centralized exchanges (Binance, Bybit, Coinbase).

Companies operating in these competitive markets seem like they have impenetrable moats, until they don’t. OpenAI had a 5+ year head start before Anthropic. FTX probably wasn’t one of the first 500 crypto exchanges to launch. Companies are extremely path-dependent and dominant players can quickly fall from prominence.

Additionally, in a Tyler Cowen-esque way, investing in these businesses provides an entry point into their peer groups and subcultures that you can later tap into for future opportunities.

3. Run an autonomous team.

The core FTX Ventures investment team was very small and had wide latitude for the majority of investments. Sign-off requirements varied depending on the size of the investment, but were much more lax even as a percentage of their overall portfolio size.

The team invested in an average of 2-3 companies per week. With fewer than eight people full-time, they clearly did not have the capacity to conduct thorough due diligence on every deal.

While FTX never structured their firm in a pod model to my knowledge, this is a natural evolution: autonomous sub-teams whose financial incentives are aligned with their performance. Pod structures are becoming increasingly common, especially among early-stage venture firms to incentivize and retain key talent.

4. Conflicts don’t matter and building a reputation for that selects for better operators.

FTX Ventures routinely invested in competitors (Circle + Paxos, Solana + Sui). Similarly, YC publicly states that they can’t promise they won’t invest in a competitor but that they will abide by basic principles of confidentiality.

Circle is now public with a $30B market cap on NASDAQ and Paxos has raised over $500M while partnering with PayPal. Solana and Sui are worth $150B and $10B, respectively. Companies were incentivized to seek FTX’s investment because (1) the obvious synergies with the parent FTX business and (2) signaling benefits.

Investing in competing portfolio companies sends the signal that choosing winners ex ante is impossible and that the investor prioritizes backing the strongest operators regardless of overlap. A new startup is incentivized to also seek investment to be among their underdog peers looking to unseat the existing dominant player.

5. Don’t overestimate your forecasting abilities; spray and pray won’t ruin your brand if it comes from a position of choice.

You’re probably much worse at market projection than you think. Engaging in intellectual masturbation where you believe you can pick and choose winners creates negative reinforcement as you don’t take action and have fewer opportunities to generate brand network effects.

Firms like a16z and Paradigm are already embracing a wider-net approach. YC is also moving in this direction, though more out of necessity than choice to deploy capital at scale. This shift changes how people perceive the rationale behind their investments.

A natural way to signal kingmaking beyond the venture brand itself is through side deals, which are becoming increasingly common. These deals provide insiders with concrete evidence of which investments partners truly believe in and want more exposure to, revealing the top-tier opportunities in a wide-ranging portfolio.

For anyone who follows the venture landscape, this is all obvious. Social capital is an increasingly liquid and compounding asset. Each company you invest in is a call option with unlimited financial and social upside.

Similarly, every blog post or tweet is a call option that could change the trajectory of your life. This may not be my best post, but lesson 5 suggests I should share it anyway.

Prediction hash: 98cac4658958b7cd2048676e2165208fba21f9d76a520fd07590cba1ecec6ec7