Jul 27, 2026 · 1h 6m · news
Will Open-Source Threaten Anthropic's Business & Do Margins Matter in a World of AI | Matt Murphy
Clips from this episode (6)
Short vertical cuts produced from the tape, captions burned in. Where a cut lands on a statement from the ledger, its card says so.
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this episode of 20VC, host Harry Stebbings interviews Matt Murphy, Partner at Menlo Ventures, discussing venture capital strategy in the AI era, Menlo's multi-billion-dollar bet on Anthropic, open-source versus proprietary foundation models, and changing fund economics.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Harry holds 24.8% of the talking time here. How this is scored →
Harry as informed peer 5.2 Guest teaching 4.6 Guest disagreement 1.4 Harry pushing back 3.9
speaking balance: gold is Harry, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 24:28 Rejection of Open-Source TAM Displacement
Matt directly pushes back on Harry's premise that open-source models will handle 96% of enterprise workloads and commoditize frontier labs, explaining how proprietary model performance directly boosts retention.
Hardest push from Harry ▶ 5:33 Mathematically Challenging the Starter CheckHarry explicitly refuses the standard narrative around starter checks, performing live fund math to demonstrate that writing $10M at $4B valuation returns a negligible fraction of a $600M fund even in massive exit scenarios.
Biggest teaching moment ▶ 7:25 Reframing Ownership Norms in Outlier VCMatt educates Harry on how venture dynamics have fundamentally changed over his 25-year career, moving away from requiring 20% ownership for moderate exits toward holding smaller stakes in generational trillion-dollar outliers.
Harry holds his own ▶ 37:43 Dissecting the Boutique Seed Fund TrapHarry demonstrates sharp venture industry insight by arguing against LP consensus, explaining why $50M seed funds writing $2M checks are structurally squeezed between multi-stage firms and agile pre-seed leads.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Harry as informed peer | Guest teaching | Guest disagreement | Harry pushing back | Why |
|---|---|---|---|---|---|---|
| 20VC Podcast Opening Title Sequence | 3 | 4 | 1 | 3 | Harry introduces Matt's recent investing streak and asks for the origin story of the Anthropic deal. Matt details how an introduction from Anjney Midha led to meeting Dario Amodei and making an initial investment out of Menlo's venture fund despite valuation friction. | |
| Reevaluating Ownership Norms in an Outlier Market | 6 | 5 | 2 | 6 | Harry explicitly pushes back on the math of writing a $10M check into a $4B valuation, calculating that an 80B exit with dilution only returns 12% of a $600M fund. Matt counters that modern VC returns are driven by outlier magnitude rather than strict 20% ownership targets. | |
| Fund Limits, Valuation Ceiling, and AI Dilution Norms | 5 | 4 | 1 | 3 | Harry asks about fund size boundaries and dilution norms across the AI stack. Matt outlines the playbook behind Menlo's $500M+ SPV follow-on into Anthropic following revenue traction and major partner validation. | |
| Managing Liquidity and Letting Outliers Compound | 4 | 4 | 2 | 3 | Harry probes Matt on taking liquidity off the table and asks about the most nerve-wracking moments during the Anthropic investment. Matt explains the difficulty of raising Menlo's first SPV and clarifies Dario Amodei's stance on secondary market SPVs. | |
| Lovable's Hypergrowth Run and AI Profit Margins | 5 | 5 | 1 | 4 | Harry cites specific valuation metrics from the Lovable deal and questions whether software gross margins still matter given high compute inference costs. Matt explains how hypergrowth AI applications transition from initial 20-30% margins up to 60-70%. | |
| Open Source Threat vs. Frontier Models and OpenRouter | 6 | 5 | 3 | 5 | Harry puts forward a scenario where open-source models perform 96% of enterprise workflows, potentially capping frontier model TAM. Matt rejects the premise, explaining that frontier models drive higher customer retention and engagement that open source cannot match. | |
| In-House Silicon and the Scale of OpenRouter | 6 | 5 | 2 | 5 | Harry cites quotes from recent show guests regarding Nebius and Fireworks to question the defensibility of model routing platforms like OpenRouter. Matt explains why developer mindshare and marketplace neutrality give OpenRouter a distinct moat. | |
| Investing in Legora and Application Defensibility | 6 | 5 | 2 | 5 | Harry raises the common bear case that base model providers like Anthropic will consume vertical legal SaaS like Legora. Matt defends Legora's defensibility by pointing to multi-party workflow complexity and context depth in law firms. | |
| Navigating the Series A Crunch with a Barbell Strategy | 6 | 4 | 1 | 5 | Harry argues that Series A is currently the worst insertion point in venture due to inflated valuations (200x ARR) without proven PMF. Matt agrees with the diagnosis and details Menlo's barbell strategy spanning seed checks and late-stage breakout winners. | |
| Boutique Seed Funds vs. Full-Stack Venture Capital | 7 | 4 | 1 | 6 | Harry delivers a strong contrarian thesis that boutique $50M seed funds will be the worst-performing vintage due to pressure from multi-stage firms writing flexible seed checks. Matt validates the shift away from rigid firm swim lanes. | |
| Disciplined Firm Sizing: Why Menlo Capped Funds at $3B | 5 | 5 | 1 | 4 | Harry questions why Menlo capped its fund at $3B rather than expanding into a mega-platform like General Catalyst or Lightspeed, and asks about partner stage flexibility. Matt defends maintaining a tight 12-partner structure to preserve alignment. | |
| Taking Calculated Moonshot Risks Outside Core Mandates | 6 | 5 | 1 | 4 | Harry points to Bessemer's spear-fish investment in Cerebras as a counter-example to strict thesis investing. Matt acknowledges that exceptional off-mandate swings are essential for VC outperformance, citing Anthropic. | |
| Why VCs Lose Deals: Relationship Depth vs. Valuation | 6 | 4 | 1 | 5 | Harry shares personal deal misses like ElevenLabs due to strict ownership targets, and references his viral debate on triple-triple-double-double growth no longer being fast enough. Matt confirms that top-decile growth expectations have completely recalibrated. | |
| Admiration for ElevenLabs and Outlier Founders | 4 | 4 | 1 | 2 | Harry asks Matt to identify a company he regrets missing out on and asks about internal partnership friction regarding the Anthropic SPV. Matt praises ElevenLabs founder Mati Staniszewski and highlights Menlo's internal alignment. | |
| Sustaining Firm Hunger After Generational Financial Wins | 6 | 5 | 2 | 4 | Harry calculates potential carry from Anthropic and asks how Menlo maintains partner drive after generational financial wins. Matt clarifies position size versus carry and emphasizes Menlo's institutional ambition in AI. | |
| Quick Fire: Mindset Shifts, Pivotal Lessons, and Plaid | 5 | 5 | 1 | 2 | Harry conducts a quick-fire round covering mindset shifts, losing Plaid early in Matt's career, preferred fund managers, and overheated sectors like Neolabs. | |
| Under-Invested Opportunities in AI Developer Tooling | 3 | 5 | 1 | 1 | Harry asks which AI areas remain under-invested. Matt identifies second-wave infrastructure, developer tooling, and observability layers above foundation models before closing the episode. |
Statements from this episode (42)
Open-source models will not displace top foundation models like Anthropic
“I think the foundation models, let's say specifically Anthropic, have such special models. This can be hard for somebody to just kind of say, I've used open source with my data. It's going to be functional and positive for some amount of what you're doing, but…”
Anthropic matched ChatGPT performance pre-launch using 1/50th of the capital
“All the benchmarks you could see that they were kind of better or at the same level of performances as ChatGPT at the time. And they'd spent like, I don't know, A 50th of the capital.”
Foundation model markets will never be dominated by one player
“These markets are never dominated by one player. There's going to be an alternative.”
Menlo Ventures pivoted its investment strategy to focus entirely on AI
“We've pivoted the firm to be all in AI.”
Menlo Ventures invested $10M in Anthropic before raising $500M+ SPV
“The first check was a little over 10 and then, you know, the, so that was kind of the starter check because the average, like I said, you try to kind of, you know, venture fund kind of have this kind of narrow window of what you invest, but then the next round…”
Stebbings: even turning 10 into 100 returns just 12% of the fund
“10 at four. I would be sitting in your partnership going, well, let's just like outcome scenario plan this. A forty billion dollar company or an eighty billion dollar company. I think you do 80. It's a 20 X with dilution. Traditional says 50%. It's a 10 X. We'…”
Small stakes in massive outliers beat high ownership in mid-sized exits
“You have to be in the big outliers to drive great returns, and you're better off being in them at a very small percent than owning a large percent of a company that exits for three to 500. Those just aren't going to move the needle.”
Menlo Ventures does not aspire to raise a $10B+ fund
“We don't have a 10 or twenty billion dollar fund, nor do we aspire to have that.”
Menlo Ventures signed Anthropic term sheet two weeks after LP meeting
“So we literally came out of that meeting and said, all right, we've got to do this. We've got to figure out a way to lead the round. And two weeks later, We signed a term sheet. We, you know, aggregated all the kind of demand from RLPs and folks we knew and, y…”
VC funds should never sell early stakes in outlier winners
“Like I think if you're a believer, I think more than ever, we're in an environment where your outliers, your winners will compound and drive fund returns. So those are certainly not the ones you want to sell from.”
Menlo's $500M+ Anthropic SPV was the firm's first ever
“Menlo had never done an SPV before. It just happened to be over five hundred million.”
Bad actors harm secondary markets by marketing fake startup access
“I think the problem is, you know, it's secondary markets, SPVs, they've just become, Too annoying and aggravating in the market to founders. And someone else is basically like, I don't want you marketing my stock. My, and I want to be the one who's figuring ou…”
Lovable scaled from zero to $300M ARR within one year
“Where you see a company go from zero to something like three hundred million in a year. I think we intercepted them around, well, we kind of got, tried to get in when they were around 30, 30 of error, but we around, we did was around one 50.”
AI compute costs cap startup gross margins at 60% to 70%
“You know, a lot of companies, just because of the cost of computing inference, it's harder to say you're going to be an 80, 90% gross margin company anymore. But, you know, great companies are, you know, 60, 70% gross margin.”
Using Anthropic drives higher retention and revenue than open-source models
“What's happening is companies see this, like, yes, I can get lower cost, but if I use Anthropic, it actually increases my customer retention. I generate more revenue. I get users to Engage with the platform more. And that is what the data is suggesting now wit…”
Enterprise AI workloads will split across proprietary and open-source models
“So I think, you know, you're going to have the combination of something like that, a family of models from Anthropic, and then a set of open source models and things that you train with your own data, and you're going to look across that whole tapestry and say…”
OpenRouter is wildly profitable and operating at unexpected scale
“Like this company wildly profitable and, you know at a scale that would probably shocked most people”
Menlo Ventures wrote a sub-$50M check for legal AI startup Legora
“The round that just happened, you know, about six months ago. It was like it was kind of sub-fifty, but in that range.”
Legal AI startup Legora plans to expand into tax and accounting
“I mean, look, Max, I guess maybe he hasn't been as public about it, but absolutely that's part of the strategy. You know, when we got to know each other and we were thinking about the round and justifying not only the current round and hopefully, you know, par…”
Series A is the worst stage due to 200x ARR valuations
“I think Series A is the worst place to be today, and my partners always hate me for this, because all Series A founders are like, great, we won't go and see them but it's the worst place to be. You have, like, one to three million in revenue, and you're at 200…”
Menlo Ventures raised its instant seed check limit to $8M
“We've got the specific seed strategy where three partners can, you know, right up to an eight million dollar check, like on the spot that used to, that number used to be three.”
Menlo Ventures has invested in approximately seven AI labs
“We have very specific strategy around Neo labs too. We're in about seven of them.”
Sub-$100M boutique seed funds will be the worst-performing VC category
“I think one of the worst performing groups in terms of venture in this vintage will actually be the small boutique seed funds, which is what every single LP that you speak to today, Matt, wants. Every LP. This is the funniest thing. Every LP wants San Francisc…”
Negative signaling from institutional seed investments is obsolete
“Everyone used to make this argument in the seed world, like, oh, there's negative signaling if you let an institution in, in there, and I think that's kind of out the window, As well, because for the right companies, like everybody's getting preempted and the …”
Multi-team VC platform expansion destroys partner alignment
“When you go full, full stack and you have like five different teams, you start doing sector, like everybody's kind of out for a pass and Sometimes I've seen this in other places where you feel like, well, I could do whatever I could do great things, but I can'…”
Menlo Ventures broke its standard investment criteria to back Anthropic
“And I'll, that's, I'll take you back to our investment in, in Anthropic. Like, same thing. It's like, This doesn't really fit. This isn't what we normally do, but wow, if this works, I mean, you got such a special founder in Dario and an amazing market.”
European founders have more grit than Bay Area founders
“There's kind of more grit to be a great entrepreneur in Europe than let's say in the Bay area where, you know, it's not that it's not incredibly hard to get into YC and just be a founder. I think in Europe, it's always been a lot harder.”
Lacking a prior relationship is the main reason VCs lose deals
“The thing that's most often is that you were late to the party, right? Like, you know, you were not intentional enough that this was a company that you wanted to be tracking and building a relationship, so you're coming in, you know, a couple weeks or a month …”
VCs rarely lose deals on valuation alone for top companies
“It's rarely just, like, straight up, you know, valuation, stuff like that. Yes, valuation can be painful. But for the right companies, you know, you do what it takes to be in.”
Rejecting 1% allocations in Deel and ElevenLabs cost Stebbings massive returns
“There have been several companies where we've had, like, one percent offered to us, Deal, 11 Labs, Starcloud, where we were like, one percent, we can't be doing that, and now I look back and all of them would have returned huge amounts of money.”
Menlo Anthology fund backed 50+ startups including OpenRouter and Whisper
“Like that's over 50 companies, somewhere between a hundred K and one million where you kind of get in a seed round and the companies that have graduated out of that have been open router, whisper, axiom math.”
Small initial check increases next-round lead likelihood by 10x
“And I would say if you get even a wedge into a company, you're 10 X more likely to be able to participate significantly in the next round or lead.”
Traditional 'triple-triple-double-double' startup growth is no longer attractive for VC
“I turned down a company the other day, Because they were going from, like, one and a half to five to 15, and there's an opportunity cost of capital to say that's very real, and the growth expectations are just very different. In other words, triple, triple, do…”
AI startups are reaching $100M revenue in one year
“When we look around and see these companies doing zero to a hundred in a year, never seen anything like it. And there's more examples of that than I can probably count right now.”
ElevenLabs' Mati Staniszewski is an outlier founder like Jeff Bezos
“The one company that I've really admired and, you know as the kind of like outlier entrepreneurs, we, in my history going back, like you look at the companies that became great, you know, when I was early days of client, there was like, you know Jeff Bezos and…”
Outlier venture investments do not require internal partnership dissent
“I don't really ascribe to this point of view where you need like a bunch of no's and there's one person who's a yes, and that leads to an outlier. I know there are examples of that, but that's not really been my experience in the firms I've been part of or wit…”
Menlo Ventures' Anthropic position is worth over $10 billion
“No, our position is, is north of that, but that's not that you can do the math on what carry usually is.”
Wealthier VCs perform better by prioritizing upside optimization
“I think richer investors make better investors because you do not worry about downside mitigation, but you focus on upside optimization.”
Market will not sustain 60 independent AI model companies
“There's no way in hell that, you know, we're going to have 60 independent model companies in addition to all the open source and everything.”