Venky Ganesan: Venture Capital Portfolios Should Be Viewed as a Basket of Options, Each Seed Investment is an Option Bet
Menlo Ventures partner Venky Ganesan stated on the 20VC podcast that venture capital portfolios should be viewed as a basket of options, with each seed investment being an option bet to observe whether a company can become an outlier.
Ganesan mentioned that seed checks buy information rights and board seats, rather than early heavy investments. Only when there is quantifiable revenue evidence confirming that a company is becoming an outlier will they scale up their positions. He also noted that Menlo is not sensitive to price at the seed stage, as the real goal is to first secure a position at the table and then increase the investment upon confirming the outlier.
He provided a dilution assumption that equity would be diluted by about 60% from the first check to exit. Based on this calculation, a 10% equity stake obtained at the seed stage may only remain around 3.5% to 4% at exit. Therefore, he is more concerned with holding a small piece of a giant company rather than a large proportion of a small company, stating he would prefer 2% of a trillion-dollar company over 20% of a hundred-million-dollar company.
The podcast mentioned that Menlo entered Anthropic in 2023 at an estimated valuation of about $4 billion, when it was still in the pre-product, pre-revenue stage, and later participated in a financing round at an estimated valuation of about $18.4 billion, with a scale of $750 million. Ganesan stated that Menlo's stake in Anthropic is less than 2%, but a single fund's allocation to the company can be up to 20% of the fund size, and it is a stepwise increase, not an early full investment.
Harry Stebbings noted in the discussion that teams are finding it increasingly difficult to see seed rounds below $100 million. Concurrent English reports indicate that common sizes for AI seed rounds have reached $10 million to $20 million. Ganesan defined the game after the market confirms the outlier as entry and position management, believing that at this point, the choice itself no longer has excess returns.
This reflects a change in event-driven allocation rules. Large funds use price-insensitive small checks at the seed stage to exchange for information and follow-on rights, concentrating capital into a few companies after confirming revenue curves. Beneficiaries are platform funds that can continuously secure follow-on amounts and validated AI companies, while seed funds that cannot follow up and can only remain at initial small equity stakes are under pressure.
Ganesan has been in the industry for about 28 years, has made the Forbes Midas List three times, and has served as the chairman of the National Venture Capital Association. His public investment path includes Palo Alto Networks, Upwork, Poshmark, and Rover, while Menlo's recent portfolio also covers Anthropic, Lovable, Legora, and Higgsfield.
Source: Public Information
ABAB AI Insight
Ganesan's approach is not a new fund manager's slogan. He first led investments in Palo Alto Networks, Upwork, and Redfin at Globespan, then applied the same "enter first, concentrate later" strategy to cybersecurity and AI at Menlo. Anthropic is an extreme example of this approach: entering at an estimated valuation of $4 billion in 2023, then participating in a $750 million financing at an estimated valuation of $18.4 billion, with a single fund allocation up to 20%, yet the final stake remains less than 2%. Failures and misses are framed similarly; he discussed the most expensive misses on the show, where the core issue is not a wrong direction but failing to exercise options after evidence appears.
Capital is not spread evenly but mobilized in two phases. The first phase uses small seed checks to buy information rights, board or observation seats, tolerating higher prices. The second phase only concentrates fund amounts into a few names after revenue data confirms outliers. Menlo's brand benefits from Anthropic are directed towards Lovable, Higgsfield, Legora, effectively using a validated leading position to secure access to subsequent projects. The motivation is specific: excess returns at the seed stage come from selection, while after outliers are public, excess returns only come from whether one can add to the position.
This logic is structurally similar to Benchmark's early position in Uber, maintaining follow-on rights in subsequent rounds, and is close to Thrive and a16z's practice of using brand to exchange for allocations during platform phases. The difference is that AI has raised seed round sizes to a common $10 million to $20 million, with even early rounds reaching $100 million, transitioning the industry from an "ownership race" to an "options and exercise" phase. If seed funds cannot put out the next large check after confirmation, initial equity will be diluted by about 60%.
Structurally, this represents a transfer of pricing power. Large funds treat seed prices as entry fees rather than valuation anchors, as true returns occur in subsequent concentrated allocations. The mechanism is that computing power and foundational models have imposed structural costs on every startup, allowing limited partners to directly buy large tech stocks in the public market, while private equity must cover costs with higher internal rates of return. Thus, seeds are no longer belief positions but statistically sufficient options, with pricing power shifting from early founders and small funds to large platforms that can continue funding after evidence appears.
ABAB News · Cognitive Laws
- Small checks buy information, large checks buy certainty.
- Selection has excess, but after confirmation, only allocation remains.
- Holding proportion is not return; exercising ability is.