20VC: a16z's $15BN Fundraise with Alex Rampell | The Best Companies Have Hostages Not Customers | The Best Founders Materialise Capital, Customers and Labour | Mid-Sized Funds with Die and The Future of Venture Capital

Rampell explains why venture capital is bifurcating into two winning categories: large generalist funds like a16z that can deploy massive capital into late-stage rounds, and small specialist funds focused on specific verticals. Mid-sized generalist funds face existential challenges as they can't compete on either dimension.

• •Companies now go public much later (Series D didn't exist in 1992), creating opportunity for larger venture funds to deploy capital in growth stages
• •LPs prefer absolute dollar returns over multiples - a 3x on $1B beats a 5x on $50M
• •Small specialist funds (like Ribbit in fintech) can win deals through deep domain expertise despite smaller check sizes
• •Mid-sized generalists lose to both large funds (on resources/network) and small specialists (on expertise)
• •The best consensus deals go to firms that can either offer unique specialization or comprehensive support at scale

Rampell outlines his framework for identifying exceptional founders, focusing on their ability to materialize three critical resources and their motivation level. He emphasizes the importance of founders who have studied the history of their space and possess a 'Count of Monte Cristo' revenge/redemption motivation.

• •Best founders can snap their fingers and get 5 people to follow them for a 50% pay cut (materializing labor)
• •Founders must be able to convince investors to give them money repeatedly (materializing capital)
• •Getting first customers is as hard as getting first employees - ability to close early deals is critical
• •Founders who study the history of their space deeply (like Patrick Collison meeting Dee Hock, founder of Visa) have massive advantages
• •The 'Count of Monte Cristo' motivation - founders driven by revenge/redemption rather than just money - are more likely to turn down acquisition offers and build generational companies

Rampell introduces his concept that 'the best companies have hostages, not customers' - explaining why sticky systems of record with high switching costs are more valuable than fast-growing but easily replaceable products. He discusses the greenfield bingo strategy of selling to new companies rather than trying to displace incumbents.

• •Companies with hostages (like Workday, NetSuite) can't be displaced even by marginally better products due to switching costs
• •Greenfield bingo strategy: sell to newly created companies who will pick the best product since they're not locked in
• •Mercury never stole a customer from SVB before it failed - they only won new company formations
• •Rate of new company creation is critical - if too slow (like hospitals for EHR), greenfield strategy fails
• •Boring, sticky products that become systems of record provide infinite option value for adding features later

Rampell discusses how the time to build competing products has compressed from years to weeks due to AI and cloud infrastructure. He explains the three types of SaaS companies in the AI era and why some will thrive while others face existential threats.

• •VisiCalc took 5 years to lose market share; in 2025, this can happen in weeks due to AI-enabled development
• •Three types of SaaS: (1) Impervious to AI with distribution advantage, (2) Existentially threatened (like Zendesk), (3) Middle ground requiring adaptation
• •Building a software product now takes 2 weeks instead of 2 years, massively increasing competitive pressure
• •Best defense is becoming a system of record with all customer data, making switching painful
• •The battle remains: will the startup acquire distribution before the incumbent acquires innovation?

Rampell reveals his belief that only 5% of current unicorns will ever be able to go public, creating a fundamental liquidity problem in venture. He discusses the dangers of massive secondaries and how they create moral hazard by misaligning incentives between founders, employees, and investors.

• •Of all unicorns, probably only 5% will ever be able to go public - a shocking statistic given the number of billion-dollar private companies
• •Many unicorns are shrinking and don't conform to Rule of 40, making them unpublishable
• •Massive founder secondaries ($50-100M+) create moral hazard by disconnecting founders from employee and investor outcomes
• •Secondaries can work if founders just turned down a $10B acquisition and remain motivated, but often lead to complacency
• •The foie gras-ing of startups: excess capital leads to doing 10 things instead of 2, breaking culture and focus

Rampell outlines the three core investment theses for a16z's apps fund: greenfield systems of record, software that replaces labor, and companies with unique walled garden data moats that even AGI cannot replicate.

• •Thesis 1: Greenfield systems of record - sticky software selling to new companies (like Rillet for procurement)
• •Thesis 2: Software replacing labor - products that do jobs people did before, enabling hyperscaling (like Eve for plaintiff attorneys handling small cases)
• •Thesis 3: Walled gardens - unique data assets that even AGI cannot access (like Vilex with Spanish legal records or Open Evidence with medical data)
• •Labor displacement software must eventually back into sticky systems of record to avoid commoditization
• •The key question for any fast-growing company: how will you get hostages and prevent customer promiscuity?

Rampell discusses the economics of venture ownership, explaining when it makes sense to accept lower ownership versus insisting on high ownership. He shares his framework of either buying 'any percent of something absolutely working' or 'high ownership of something that could work.'

• •The framework: buy any percent of something absolutely working (like Facebook at Series B) OR high ownership of something that could work
• •If you win 100% of deals, you're probably not testing the efficient frontier on ownership - you may be overpaying
• •Series B trap: companies that just added infrastructure/scaffolding without metric improvement - same risk, half the ownership
• •Winning deals at 5% ownership only works for true rocket ships with unprecedented growth rates
• •Must balance win rate against ownership - can't optimize for 100% win rate if it means 5% checks everywhere

Rampell introduces the concept of 'founder-capital fit' - ensuring founders can handle large amounts of capital without losing focus or creating moral hazard. He discusses the Rillet case where he invested in Series B just 60 days after Series A.

• •Founder-capital fit is as important as product-market fit - some founders can handle $1B without losing focus, others cannot
• •Infinite capital removes the forcing function of hard decisions - founders need constraints to stay sharp
• •The worst option is often making no choice at all when faced with two bad options - capital can enable this paralysis
• •Rillet Series B case: sometimes you must pay up for winners even if it's expensive, if you believe in the founder's motivation
• •Key question: if I give you a billion dollars, what will you do with it? 99/100 times the answer reveals problems

Rampell shares tactical advice on selling companies, emphasizing that successful M&A requires years of relationship building with the right people at potential acquirers - not corp dev, but the SVPs who will actually manage the acquired team.

• •Spend 5-10% of CEO time as a 'background process' building relationships with 3-4 potential acquirers years in advance
• •Never approach corp dev directly - they execute transactions but don't make buying decisions
• •Build relationships with SVPs who have budget holes or need revenue growth for bonuses - understand internal mechanics
• •Best time to sell is when company is doing great, but this rarely intersects with when you need to sell
• •Use partnership discussions as cover for relationship building - if the partnership is valuable enough, acquisition often follows naturally

Rampell shares his vision for venture capital's future, predicting that technology will continue eating more of the world as AI enables software to replace labor in entirely new categories. He sees vertical SaaS expanding dramatically beyond current markets.

• •The five biggest companies on Earth are now technology companies (were banks 20 years ago, oil companies before that)
• •AI is enabling software to do things it couldn't before, expanding addressable markets 100x
• •Vertical SaaS proved the model with Toast at $20B valuation - expect many more category winners
• •If robotics actually works, the market expansion could be another 100x beyond current software opportunities
• •Technology will continue creating enduring value as it pervades every industry and function