20VC: Is Seed Investing Dead Without a $1BN Fund? | Does Ownership and Price Matter When Companies Can Be $1TRN Exits | Are AI Revenue Numbers Real and What to Watch Out For with Venky Ganesan, Menlo Ventures

Summary

Venky Ganesan says Menlo is aggressively positioning for AI and treating each seed investment as an option to discover outliers, with position sizing increased only after quantitative proof emerges. He argues seed is especially hard in core AI because large funds use it as an access point and may be indifferent to price, while founders and investors increasingly optimize metrics like ARR and retention once those become highly weighted. Ganesan is skeptical of revenue run-rate style signals, noting that accounting and financing structures can game what looks like growth, and he thinks reflexive markups, leverage, and debt are what eventually break cycles. On ownership, he says it still matters, but relative to the scale of the outcome: a small slice of a trillion-dollar company is better than concentrated ownership in a small one. He also argues that in AI venture, IRR matters more than ever because investors are effectively paying a “tax” to Nvidia, hyperscalers, and foundation model providers, and venture must compete with public-market index returns.

Notes

• Menlo is pursuing maximal AI exposure and framing the strategy as an all-out hunt for category-defining winners.
• Seed investing is modeled as an option on an outlier, not a commitment to size up immediately.
• In core AI, seed is hard because large funds treat seed as an access check, not a price-sensitive bet.
• Menlo says some large funds are indifferent to seed valuation because they are using the round as an option check.
• The point of participating early is to buy a seat at the table and later size up if the company becomes an outlier.
• He says metrics that matter to investors will be gamed, including SaaS-style retention measures.
• He describes markup-driven growth as a reflexive loop that can accelerate capital, talent, and attention until the cycle stops.
• He thinks cycles usually crack first through debt defaults, not equity write-downs.
• Tranche financing began as a useful distinction between value-add capital and pure capital, but has become generic and overused.
• He will still participate in later tranches when the company and founders are exceptional.
• Ego should not determine capital allocation; the objective is to maximize returns for LPs.
• He argues venture is a business of humility and execution, not personalities.
• Menlo will pay up when necessary and wants founders to call the firm.
• Price premiums can reflect either deal-winning aggression or a bigger view of TAM.
• He regrets missing a Sean Parker opportunity around Plaxo and says seed rounds then were tiny by today’s standards.
• He looks for founders who can explain complex ideas simply and show real insight.
• Company DNA is set early by the founding team and how they understand each other’s strengths and weaknesses.
• He uses a three-best-friends question to test founder self-awareness and align references.
• Ownership still matters, but only relative to outcome size.
• Menlo owns less than 2% of Anthropic and says ownership cannot be viewed in isolation.
• The right game is to buy ownership early, then size up only once the company proves it is an outlier.
• Once a company is known to be an outlier, the game becomes access and sizing, not selection.
• Menlo says it would not usually put more than 20% of a fund into one company, but it has done so once.
• Anthropic was the exception where Menlo hit 20% in a single company.
• Position sizing should follow new data, and even 10% ownership is hard in the current market.
• Current venture excitement is driven by very fast growth, but he doubts that regime lasts forever.
• He is excited by a company that went from $250 million to $10 billion in 10 weeks, but wants to understand the underlying data and growth cost.
• AI-era strategic acquisition prices look rational only because giant buyers can afford them, but investors should not assume rescue valuations.
• The AI acquisition market feels like late dot-com, where buyers paid for teams even without product or revenue.
• Menlo assumes big dilution between seed and exit because financing rounds and option pools expand over time.
• Dilution is mainly a function of time horizon, with fast exits and fast step-ups helping IRR.
• In AI, venture has to focus on IRR because every company pays rents to Nvidia, hyperscalers, and model providers.
• Private-market returns now have to clear public-market alternatives by a wide margin.
• Menlo prefers a two-way commitment with entrepreneurs and says that influenced its decision not to invest in OpenAI.
• AI forces faster recycling of funds because capital needs and growth rates are much higher than in older software eras.
• Menlo 8 was the only fund in the firm’s 50-year history not to return capital, because it was invested over a 10-month window in 2000-2001.
• Menlo’s larger early-2000s funds underperformed, and LPs departed.
• LPs want DPI, and showing real distributions helps fundraising.
• At very large marks, the question is whether to harvest some gains without fully exiting.
• Taking 10% to 15% off the table can let investors and founders stay aligned and go longer.
• He generally prefers not to sell out completely unless the company itself is being sold.
• Founder mode is a behavior pattern, not a title reserved only for the original founder.
• More AI-era M&A should emerge because regulators may allow it and large acquirers can pay relatively modest sums.
• Mark Zuckerberg is cited as a strong capital allocator, and capital allocation is treated as the key founder skill at scale.
• Capital allocation, in his view, subsumes product vision because it forces choices about where to deploy resources.
• He says Snap has strong product vision but poor shareholder outcomes over the past seven to eight years.
• Money reveals character rather than changing it.
• He attributes California housing inflation mainly to supply constraints and NIMBYism.
• He praises Brian Armstrong for taking a principled stance against political activism inside Coinbase.
• He says he is less willing to say inauthentic things to win deals now, even if that means losing some.
• Wealth reduces fear and allows investors to take more concentrated risk.
• $30 million to $100 million funds are at a disadvantage versus larger chip stacks.
• A $200 million fund can still compete if it gets in early and selective.
• There is no magic strategy in venture; success comes from hustle and grit.
• Venture-backed SaaS turnarounds are harder than PE because no one owns enough to control them.