The Twenty Minute VC · 5 Oct 2026 · From the week of 5 October
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
Listen on Apple Podcasts Show on Spotify Episode page
In brief
Harry Stebbings interviews Venky Ganesan, a partner at Menlo Ventures (backer of Anthropic, Higgsfield and Legora), about how venture investing works in the current AI boom. They cover seed checks as cheap options followed by data-driven position sizing, gamed revenue metrics, tranche rounds and kingmaking, ownership versus outcome size, dilution and IRR benchmarks, the risk of assuming big tech will buy a company, when to sell winners, LP dynamics and the outlook for emerging managers and zombie SaaS. Ganesan's central argument is that investors have to keep playing in a frothy market without timing it. He says they should treat early checks as options, ladder up concentration only on quantitative evidence, keep ego out of pricing and syndicate decisions, and judge returns by IRR against a public-market index that already owns the Magnificent Seven.
For founders
- Ganesan expects a VC who owns 10% at seed to hold about 3.5-4% at exit, roughly 60% dilution, though fast-scaling companies with quick exits suffer much less.
- Ganesan said the right time for investors to take some chips off the table is when the founder is also taking some off. He noted that 2021 SaaS holders who had sold even 10-15% would have locked in gains, and he presented selling as something that lets both sides keep holding.
- Ganesan warned that a large-acquirer floor ('worst case big tech buys us') is a dangerous thesis, because acquirers in structured deals can hire the founders and leave investors with nothing.
- Ganesan said investors weight metrics like NRR and annualized run rates heavily and so those metrics get gamed. He predicted the next few years will expose founders' 'accounting creativity', and said he wants founders focused on terminal value rather than markups.
- For founders at scale, Ganesan would choose capital allocation over product vision, because good capital allocation already includes picking the product direction.
For revenue leaders
- Ganesan gave an example of net revenue retention being gamed: land a small $10 contract and expand to $50 a week later rather than sign $100 up front, and NRR reads as 500% instead of 100%.
- Ganesan noted that a fast rise in valuation sharply cuts the equity cost of senior hires, which he illustrated by contrasting 2% of a $200M company with RSUs worth $20M at a $2B company.
What was said 32, most useful first
Large funds, Menlo included, are somewhat indifferent to seed valuations because they treat seed checks as the price of a seat for later, larger investment.
Ganesan admitted that Menlo is 'trying to buy ourselves a seat at the table' at seed, and that the real goal is to size up later with large amounts of capital. He named this as one of two forces reshaping seed, the other being much larger round sizes; AI application seeds that used to be $3-5M are now $10-20M.
“the cost of buying that seat of the table is somewhat indifferent to at the seed stage because our real goal is to size up”
Listen on Apple Podcasts Fundraising & investors Link to this
Any metric investors weight heavily will be gamed, and he expects this cycle to expose founders' accounting creativity.
Ganesan pointed to the SaaS era, when net revenue retention could be inflated by landing a $10 deal and expanding to $50 a week later rather than signing $100 up front, which turns 100% NRR into 500%. He referenced the economist's idea of the 'bezzle' rising in booms, but was unsure who coined it. He said the real filter is whether founders and investors are focused on terminal value or on markups.
“Once a metric is measured, it can be gamed. And that happens.”
Ganesan warned that assuming a large strategic acquirer will buy a company at the preferred stack is a dangerous downside-protection thesis.
Ganesan noted current large deals, such as AMD's $8.5B acquisition and Nvidia buying Hugging Face for $14B, and compared them to the dot-com era, when acquirers such as Lucent paid billions in stock for product-less teams (Lucent paid $4.5B for Chromatis). He said that thesis 'didn't quite work out' after March 2000. He also noted that acquirers have no reason to protect investors and can hire founders through structured deals that leave investors out.
“I just wouldn't take the mindset, oh, some large strategy is going to buy my company for the preference stack because they don't care about the investors.”
Menlo assumes about 60% dilution from its first check to exit, so 10% at seed becomes roughly 3.5-4%.
Ganesan said the dilution comes from both financings and option pool expansions. When Harry pointed to fast-scaling companies like OpenRouter taking far less dilution, Ganesan said the main driver is time horizon: companies that compound value fast and exit quickly take less dilution. Long horizons both raise dilution and hurt IRR.
“we assume by the time we sell likes at the company, if we own 10%, we would have three and a half to 4%. We expect 60% dilution from the point of our first check”
Listen on Apple Podcasts Fundraising & investors Link to this
Fast deployment is defensible because AI companies need huge capital, but vintage concentration has cost Menlo before: Menlo 8, deployed in 10 months, is its only fund in 50 years not to return capital.
Ganesan said LPs cannot have both smaller funds and slower returns to market when AI growth demands capital; Google probably raised less than $50M privately, but AI companies need compute. Menlo 8 was invested over 10 months in 2000-2001, and its subsequent $1.5B (2001) and $1.2B (2004) funds also underperformed. He said GPs should not follow dogmatic rules, but should weigh time diversification and communicate transparently with LPs.
“we only had one fund that's not returned capital, just Menlo 8, which was invested in a 10-month period between 2000 and 2001.”
Listen on Apple Podcasts Fundraising & investors Link to this
Ganesan recommends that investors sitting on a 30-50x return take some off the table, ideally when the founder is selling a secondary.
Ganesan said that if holders of 2021 SaaS positions had sold 10-15%, even small amounts, they would have locked in gains. He tells founders that taking chips off makes both founders and investors more willing to 'go long', which keeps them aligned. He said it is not a fund-size issue, and illustrated this with Harry's 40x position. He personally learned the lesson from watching a $5,000 IPO allocation in Avanex rise to $200,000 and then sell for about $8,000-9,000 after a 90% drop.
“If people had taken 10%, 15% off the table, even if it's small, it locks in, allows you to go long.”
Listen on Apple Podcasts Fundraising & investors Link to this
Venture-backed 2021-vintage SaaS companies are in a worse position than PE-owned ones because no shareholder has enough control to act.
Ganesan said PE owners at least have majority control and can restructure. Venture-backed SaaS companies that raised at high multiples have become 'zombies' that nobody owns enough of to fix. For that class, he said the best outcome is being bought by Bending Spoons ('getting spooned', which means getting capital back) and the worst is zero.
“You have a lot of zombie SaaS companies where nobody owns enough to be able to do anything.”
Ganesan advises LPs to judge venture firms by which partners leading AI founders respect, since fund performance lags five to seven years.
Ganesan called past returns a rear-view mirror. His suggested test is to call founders of successful AI companies and ask which partners they respect, even partners they did not take money from. He said that if a firm has none of those partners, that is the warning sign.
“call a bunch of entrepreneurs or successful AI companies and ask them who are the partners they respect. And my aim is they didn't take the money.”
Listen on Apple Podcasts Fundraising & investors Link to this
Venky Ganesan treats each seed investment as an option and sizes up only when revenue and quantifiable metrics show a company is an outlier.
Ganesan said a fund needs enough 'at bats' so that it is likely to hold an outlier, and should then concentrate capital only on companies already proven by quantitative evidence. He presented portfolio composition and position sizing as the way to manage cycle risk without trying to time the market.
“Each seed investment is an option bet. You're buying an option to see if it's an outlier... then you only position size up on the things that are already proven?”
Listen on Apple Podcasts Fundraising & investors Link to this
Investors should keep investing during booms rather than step out, because firms that exited the dot-com boom early missed its best years.
Ganesan described smart firms that invested in 1993-94, stepped out in 1996-97, missed 1997-99, and stepped back in around 2000, to their LPs' dismay. He said that rather than sitting out, investors can be more selective and lean on portfolio composition and position sizing. He also cautioned against setting long-term strategy from a snapshot of a disorienting market that 'changes quickly'.
“timing markets is really, really hard. I think you have to play the game, but I think you can play the game differently.”
Ganesan described kingmaking through rapid successive markups as Soros-style reflexivity that works while revenue is real but eventually stops.
In Ganesan's account, fast revenue growth leads to a quick markup, which brings more capital, press and talent, which drives faster growth and another markup. He warned that copycat founders and investors come to believe the markup itself is the secret, and that the dynamic then drives the market cycle until it stops. He added that no one knows how or when it will stop, and that until it does 'a lot of people can look very smart'.
“one thing we know from Soros is that all reflexivity will eventually stop. We just don't know how and when.”
Listen on Apple Podcasts Fundraising & investors Link to this
He thinks a market cycle usually cracks first with a major debt default, not with equity losses.
Ganesan reasoned that equity losses are simply written down, whereas debt holders expect repayment, so leverage is what breaks cycles. He said that with leverage, being right is not enough: you also have to get the timing right.
“I think usually the first sign comes with some major debt default. Generally, equity is never the reason why these things crack”
Tranched rounds began as a sound way to separate value-adding capital from cheap capital, but have spread to companies regardless of quality.
Ganesan described the original logic as taking 'build with me' money at a low valuation and then raising pure capital at a higher valuation. He said that, as in every cycle, the innovators were followed by imitators and then 'idiots', so the technique is no longer tied to company quality.
“in every cycle you get the innovators, then you get the imitators, and then you eventually get the idiots.”
Listen on Apple Podcasts Fundraising & investors Link to this
Menlo will invest in a later, higher-priced tranche if the company and founders are special, because ego should not drive investment decisions.
Asked whether it is insulting to pay double what another lead paid days earlier, Ganesan said he does not care what others invested and that his 'only ego is to make money for my investors'. He admitted his own ego has got in the way, through valuation negotiations, syndicate positioning, or turning down an allocation he felt was too small. He called ego-driven decisions the biggest mistakes.
“If there's an opportunity to make money on investment, we should do it. The rest of this is all noise.”
Listen on Apple Podcasts Fundraising & investors Link to this
Because venture returns are asymmetric, the deals a VC passes on are costlier mistakes than the deals it does.
Ganesan said some investors pay up only to win, while others pay up because they see a bigger TAM, and in that case the price reflects a bigger opportunity rather than overpaying. Because losses are capped at the dollars invested and wins can return 10x or more, he said sins of omission outweigh sins of commission.
“the most expensive mistakes venture capitalists make are the deals they passed, not the deals they did.”
Listen on Apple Podcasts Fundraising & investors Link to this
Ganesan looks for founders who can explain very complex concepts simply and show real insight, a lesson he draws from passing on Sean Parker's early pitch.
As a young Plaxo board member, Ganesan declined even to take a meeting when Sean Parker, who had just been removed from that board, invited him to get involved with a 'college dropout'. Ganesan said he probably could have written a $50,000 check into a roughly million-dollar seed round. He said Parker understood virality, network effects and human behavior and could boil them down simply, and that this is now his rule of thumb for founders.
“I'm always looking for people who are incredibly good at communicating very complex concepts in a simple manner and just have insight.”
Listen on Apple Podcasts Fundraising & investors Link to this
Ganesan probes what brought a founding team together because he believes the founders set the company's DNA and future hiring decisions.
Ganesan asks why these particular people decided they should be the ones to build the company, and how they see each other's strengths and weaknesses. He treats the answers as a 'precog' for how the founders will make decisions and build the rest of the team.
“I think the company you build is a team you build. And so so much of the DNA of a company is set by its founding team.”
Listen on Apple Podcasts Hiring & team building Link to this
Ganesan asks founders how their five best friends would describe them in three words, then checks whether references match that self-assessment.
Ganesan said describing yourself through friends' eyes externalizes the question and reveals self-awareness. He then runs references to see whether they match. In his view, founders who know their weaknesses can manage them; the ones blind to their weaknesses usually have trouble.
“I'm trying to see if the references match someone's self-awareness... It's the people who are blind to the weaknesses that usually have challenges.”
Listen on Apple Podcasts Hiring & team building Link to this
Ganesan would rather own 2% of a trillion-dollar company than 20% of a $100M company, but says early ownership still matters before an outlier is obvious.
Menlo owns less than 2% of Anthropic. Ganesan said that before a company is clearly an outlier you are in the 'ownership game', and afterwards it becomes a position-sizing and access game with 'no alpha'. His example was Higgsfield, where his partner Amy got 15% for a $5M check and Menlo then also sized up.
“I'd rather take 2% of a trillion-dollar company than 20% of a hundred-million-dollar company”
Listen on Apple Podcasts Fundraising & investors Link to this
Early ownership barely matters because the early check mainly buys information to size up later.
Harry Stebbings framed the early check as a way to gain information, after which success becomes a capital-concentration game. Ganesan disagreed in part: being in the company beats not being in it, but being in with meaningful ownership drives real returns, since once an outlier is known it is purely an access and sizing game.
“For actually, it doesn't matter if you have ownership in the first place. What you're buying is the information to size up.”
Listen on Apple Podcasts Fundraising & investors Link to this
Menlo caps single-company exposure at about 20% of a fund, reached that level once with Anthropic, and builds such positions by laddering up on new data.
Ganesan said Menlo would not put more than 20% of a fund into one company and has done so only with Anthropic. He distinguished making a 20% bet at the start of a fund, which takes far more risk, from laddering up to 20% as new data arrives. He said that with ownership falling ('even 10% is hard') investors should size up in proportion to the outlier opportunity.
“did you put 20% of the fund of a company in one check at the beginning of the fund? Or did you ladder up to 20% on the base of new data? Obviously much better to ladder up”
Listen on Apple Podcasts Fundraising & investors Link to this
Meaningful ownership is insurance, so a fund that misses the outliers can still return from mid-size outcomes.
Ganesan said that with low ownership, a fund is playing a game with only grand slams and strikeouts, and that a fund that strikes out repeatedly will be 'a tough fund'. Ownership lets a triple still move the needle.
“part of getting ownership is giving some insurance for you that if you missed out on the outlier, the sort of midsize outcomes can still move the needle for you.”
Listen on Apple Podcasts Fundraising & investors Link to this
A fast rise in valuation sharply reduces the equity a company gives away to hire senior executives.
Ganesan said business velocity matters a great deal to VCs, partly because of this effect. A $200M company might give 2% to hire a senior exec, he said, while a company that quickly reaches $2B can give the same person $20M in RSUs, which he put at 0.1%.
“You quickly become a $2 billion company, you don't need to get, you're going to give RSUs and you give the same person $20 million”
Listen on Apple Podcasts Hiring & team building Link to this
VCs must now optimize for IRR and beat a no-fee public index by about 1,000 basis points, because AI startups all pay the Magnificent Seven.
Ganesan said that 28 years ago investors focused on cash-on-cash returns and IRR took care of itself. Today, every venture company pays Nvidia, a hyperscaler and probably a foundation model provider, all of which are or will be investable in no-fee, no-carry index funds. To justify private-market capital, he said, a venture fund must beat that by roughly a thousand basis points.
“there's no way for Venture to be successful in today's era without the Mag 7 participating in everything you're doing.”
Menlo does not back competing companies once it takes a board seat and writes a big check, which is why among the AI labs it stayed only with Anthropic.
Ganesan described this as a cultural choice shared by his partners. A large commitment to a founder should be a two-way street, he said, while small passive seed checks are a different matter. He acknowledged that other firms back several players in the same space and said each firm must act on its own values.
“when we make a big commitment to the founder, we think of as a two-way street. They come into us, we come into them.”
Listen on Apple Podcasts Fundraising & investors Link to this
LPs keep committing to AI venture because their large software-heavy private equity books are exposed to AI disruption, but they want DPI first.
Ganesan said many LPs hold three to four times more private equity than venture, and much of that PE is in software that AI directly affects, so AI venture serves as a hedge. He said their chief complaint is too much TVPI and not enough DPI. Managers who have delivered DPI and can credibly claim a position in the AI economy can raise money.
“If you want to hedge against your private equity portfolio, you've got to be in the AI economy.”
Listen on Apple Podcasts Fundraising & investors Link to this
Ganesan rejects writing a company to zero when its founder leaves, arguing that founder mode is a way of operating, not a person.
Responding to Jason Lemkin's rule of writing such companies to zero, Ganesan said the question is who replaced the founder. He cited Nikesh Arora, Jeetu Patel at Cisco and Frank Slootman as hired executives who operated like founders.
“Founder mode is a mode of... It's not tied to anyone personally. I think anyone can be a founder in terms of working in a founder mode.”
Ganesan expects more large tech M&A, driven by a permissive regulatory window, competitive pressure, high acquirer equity prices and the urgency to catch up in AI.
Ganesan said a backlog of M&A held up under a different regulatory regime is now moving, and that this window may not last. In his view, each deal forces rival responses, as when AMD's purchase raises the question of whether Nvidia must react. He added that richly valued acquirers can do large deals cheaply relative to their market caps.
“every acquisition forces a bunch of comparative dynamics we have to consider.”
Ganesan would choose capital allocation over product vision as the key skill for a founder at scale.
Ganesan argued that capital allocation includes product vision, because allocating capital means choosing product direction, while the reverse is not true. He cited Mark Zuckerberg's Instagram ($1B) and Onavo ($400M) acquisitions as strong allocation. He contrasted this with Snap, run by what he called a product genius, whose shareholders he said have not been rewarded for seven or eight years.
“capital allocation by itself also captures product visionary because you're allocating the capital to the things that matter... The other way is not true.”
$30-100M funds are a tough place to be today, since they play against much larger chip stacks and only exceptional managers succeed.
Ganesan compared large funds to the player with the most chips at a poker table, who can see more cards. He cited Sarah Guo (whose fund he thought was about $200M) and Box Group as exceptions that should not be extrapolated from. He said there is no magic strategy, and noted Guo telling LPs her strategy is simply to work harder.
“in general, those are tough places to be unless you're exceptional... There's no magic strategy.”
Listen on Apple Podcasts Fundraising & investors Link to this
Founders underestimate how little venture investors are excited by steady growth like $1M to $4M to $8M to $16M ARR, and Ganesan agreed but called it a snapshot in time.
Ganesan explained that VCs are seeing companies go from 1 to 10 to 50 to 100 in about three years, and some go from 0 to $1B in 18 months. He said he does not think this continues forever, and urged a long-horizon view.
“going from one to four million and then four to eight million and then eight to 16 and then banking five years time, we're going to to get Venture excited today.”
Listen on Apple Podcasts Fundraising & investors Link to this
Money does not change people but reveals them, so large payouts won't demotivate A players.
Asked about sales reps leaving AI labs with $30-40M, Ganesan said truly motivated people stay motivated with money, while those only acting motivated opt out when the money arrives. For A players, he said, money is a way of keeping score and the game is what they love.
“Money doesn't change people. It reveals them.”
Listen on Apple Podcasts Hiring & team building Link to this