Jason Lemkin

1. Key Themes

Pro rata is an option with a real price, and everyone's position is conflicted

The Valar Atomics dispute exposed that nearly every public opinion on pro rata is self-interested. Jason Lemkin recalled the only clean framing, from Keith Rabois: pro rata is core to large-fund economics, so removing it means paying less upfront. Lemkin: "He's like, 'I'd just have to pay a much lower price then, worst case,' right? 'If I'm investing early with an expectation I can deploy a lot later, I'm comfortable overpaying in the early round for my pro rata. I'd have to pay like half the price I'm paying otherwise. Are you comfortable with that founders?' And the crickets." 00:15:16 Rory O'Driscoll added the tell: "all the people, even the megafunds who are coming in saying, 'You guys should waive your rights,' ain't fixing to waive theirs, which should be a clue about the real value thereof." 00:16:18 Lemkin on the biased discourse: "I've never seen more crap, more selfish takes on Twitter that are seen as generous like, you know, 'I'm only a pure seed investor, I don't need pro rata.'" 00:15:16

Pragmatic wisdom for investors in a dispute: raise it once, then let go

Both Lemkin and O'Driscoll converged on "intellectually right, pragmatically wrong." O'Driscoll: "Whenever you start an argument... you forget to ask yourself, 'What are you going to win?' First of all, you've waived your pro rata rights... you're suing for the right to get 10 or 15 million bucks in at $6 billion pre, when you already own 5% of the company, which is worth $300 million. What are you doing here?" 00:12:22 Lemkin: "Raise it once and let it go... It's not your company. It is not your cap table." 00:13:44

Meta is back at the frontier, but the real prize is the consumer product

Both hosts discounted the benchmark (no API, internal use only) but saw the product implication. Lemkin: "I wasn't that impressed with the model I can't use... The fact that... it just going to make the Muse app better, though, is super interesting... there's a material chance Meta has uh just a monster AI product on its hands. The amount of investment into it is so high, the capabilities are so high, the autonomy is so high." 00:19:30 O'Driscoll noted the strategic shift: "They were competitive early on with Open Weight, then they kind of faded back way to the back of the pack, and now they're up front again, notably with closed-source models. So we've decided that if we're going to spend $100 billion, we'll at least be able to charge for the damn thing." 00:18:02

NVIDIA vertically integrating into open models is a defensive play against its own customers

O'Driscoll framed the economics of the rumored Reflection deal: "if you're in the chip business... some of your biggest customers are closed-source model companies... you love open-source model and only one chip provider, just like on the other side of the table the model companies love closed-source models and everyone having their own chip provider... we lend each other $100 billion or so, but we're watching you closely, and we want to be able to fight back if you get margin-greedy." 00:20:39 Lemkin projected the endgame: "This is a software company, not just a hardware company... in 12 or 18 months, why would you want to use any other sort of open-weight open model other than NVIDIA's?" 00:23:06

Decacorn production is without historical precedent

O'Driscoll: "We roughly generated 20, plus or minus, decacorns per decade for the last five decades. And we generated, what did you say, 42 decacorns in less than a full year of this decade." 00:00:07 He added the scale: "130 companies in total worth 10 billion or more since the dawn of... American capitalism, and a further 42 this year." 00:32:30 Their conclusion: "either this is the most economically transformative event since the Industrial Revolution, or we're getting a wee bit ahead of ourselves." 00:33:13

The liquidation preference is the mental model that justifies stretched prices

O'Driscoll: "once you have a liq pref and you believe it exists, you can say to yourself... it's a 1x plus option value. And the thing right now is, option values have high value, because there's a lot of vol." 00:33:43 Harry pushed back that "not everything has a downside protection of being acquired," and O'Driscoll conceded: "it'll be true until it's not." 00:30:38 Lemkin's counter: today's prefs in Jev-class assets beat "a thousand liquidation preferences in a thousand 2021 unicorns." 00:31:03

Price cuts are finally producing real deflation in AI COGS, led by speed as much as price

Lemkin: "These models, Haiku 55 and Sonnet 55... are... the first thing that's actually been deflationary with LLMs and AI... None of our portfolio companies are seeing their COGS fall in AI, are they? Not really." 00:46:01 On speed, he added: "Sonnet 55 is twice as fast, 10% cheaper, not that much, and better... Speed is also deflationary because you can do more more quickly." 00:47:53 O'Driscoll set the bar for the bet: "when you cut prices by 90%, 'We're going to use more' doesn't cut it. We got to use 10 times more. Otherwise, demand elasticity is less than 1 and your actual revenue goes down." 00:50:06

The SaaSpocalypse isn't over; it just bifurcated by scale

Lemkin: "I think it only passed for those with massive scale. I think everyone without massive scale is living a worse SaaSpocalypse than in March or April, because no one wants to buy software." 01:01:51 His IPO corollary: "You might need a billion and a half or two billion of ARR as a software to have a safe IPO." 01:02:33 O'Driscoll explained the structure: "it's a mature, consolidating industry now... If you're already at 40 billion, you're money good... If you're at 100 million and you're trying to get to 300 million in a SaaS product, it's really hard today." 01:02:05

Physical-world data is booming ahead of robotics deployment

Harry pointed to multiple data providers above $100M in revenue (Mecha, MicroAGI, others). O'Driscoll flagged the risk: "I saw GXO, which is a logistics company... '42 POCs of humanoids, zero deployed.'... there's a deployment issue in robotics, which will filter back upstream to ultimately spend on data business." 00:42:15 Lemkin offered a demand datapoint: a founder "did $30 million in 60 days. And he said... 'We have $300 million of demand... we're turning away 270 out of 30 cuz we don't want to do it.'" 00:43:30

AI-for-science may be the next value-capture model beyond labor replacement

O'Driscoll on Isomorphic Labs: "here the value capture is not replace labor. It's actually make science discoveries... when you create the $200 billion drug, you want to get 10% of that on a royalty basis." 00:37:26 Lemkin updated his own priors: "the medical evidence for GLP-1s is so profound... I thought it was wackos and Bryan Johnsons and people selling crap." 00:39:23

2. Contrarian Perspectives

"Hiring a top litigation firm is the wrong move even when you feel wronged"

The crowd on X sided with the aggrieved seed investor, but Lemkin and O'Driscoll thought both sides erred and that the investor's best move was a letter and then silence. Lemkin: "Maybe you do send a letter, okay? And maybe you don't hire Quinn Emanuel, maybe you hire Stebbings and O'Driscoll, because they're not quite as intimidating... You just got to get in line and move on." 00:08:12 He also noted the mistaken narrative: "immediately everyone was all over it saying that Day One was suing their portfolio company, which was completely wrong." 00:05:49

"Take the NVIDIA bid, because saying no is itself a threat"

Lemkin gave the founder's-seat view of being courted by a dominant platform: "It's very stressful to say no to NVIDIA knowing that they're going to directly compete with you and promote your competitor... it's a 'build, buy, or partner' sort of threaten-with-a-smile message." 00:22:24 O'Driscoll: "This might be a moment in time that doesn't come back." 00:21:57

"Focused small teams beat big-company parity on classifier-style AI products"

Counter to the consensus that Cloudflare, Amazon, and OpenAI products erase startup moats, Lemkin argued maintenance intensity wins: "Will the Cloudflare team keep putting as much effort in as a standalone team?... I think as VCs, we have to make this stressful investment, that a core team that is very strong out of OpenAI that is doing nothing else will pull ahead from the seeming parity you can achieve today." 00:27:29 He added the use-case framing: "Jev is super fast and super cheap, okay? And it is accurate enough to use for maybe 10% of your AI use cases. If they make it hyper-accurate, it changes the game." 00:29:08

"Selling to a layoff-machine acquirer may be worse for employees than a clean deal, but Lemkin wouldn't do it"

Lemkin took the unfashionable ethical line: "There's no way I would sell my company to have 80% of my team laid off. I just would never do it." 00:52:49 O'Driscoll countered that retention deteriorates anyway: "the death march will get old pretty quickly... sometimes resolution is worth it even for the employees," recommending founders "cut your deal on your employee economics upfront." 00:54:26

"Early-exit SaaS wins that make a small fund don't indicate the market is healthy"

Even with TechWolf selling to SAP and Intercom's AI-native reboot, Lemkin argued these don't rebut the scale problem: "Kudos to the Intercom for building an AI-native reboot of the company to nine figures in revenue, but that's very different than building to a billion growing 50%." 01:04:14 O'Driscoll said the real issue is exit capacity: Big Tech "can only buy so many at a time... they get to go hoover it up at the rate of 10 a year." 01:05:00

3. Companies Identified

Valar Atomics

Nuclear startup. Raised a billion-dollar Series B at a $6 billion valuation led by Sequoia. Mentioned as the subject of the pro rata dispute with seed investor Day One Ventures. Harry: "Day One Ventures put about $1 million into the seed and $15 million into the A. They hold about 5%." 00:03:46 Lemkin clarified it's Valar seeking declaratory judgment: "it's Valar that wants declaratory judgment to move on. They just want an answer." 00:05:49

Day One Ventures

Early investor in Valar Atomics; roughly $450M total fund per Lemkin. Mentioned as the party asserting spring-back pro rata rights. Lemkin: "even though they're a $450 million total fund, they have more funds, whether it's SPV or whatever. So this annoyed them, they wanted to do their pro rata." 00:06:42

Sequoia Capital

Led the Valar Series B, backs Reflection and participates in Jev and Mecha rounds. Mentioned for its brand power and aggressive terms. Lemkin: "Sequoia came in very aggressively and said, 'We're going to make you a great offer. We're going to make you a fast offer. It's clean, it's simple... everyone has to waive their pro rata.'" 00:05:49 Also: "the Sequoia brand is super powerful." 00:10:45

Index Ventures

Named alongside Sequoia as a firm that takes whole rounds with waived pro rata. Lemkin: "just like Sequoia and Index and others do, everyone has to waive their pro rata so that we can get our ownership target." 00:05:49

Meta (Muse Wonder)

Released a new frontier model, just shy of Opus 5.5 on an independent index, pursuing closed-source and consumer AI. Lemkin: "There's a material chance Meta has uh just a monster AI product on its hands." 00:00:00

Scale AI

Alex Wang's company, acquired in the Meta deal. Lemkin: "To Rory's point, this may be have been a great deal for Scale. It may turn out... the next WhatsApp or Instagram." 00:18:43 Also referenced as a data-vendor comparable. 00:44:39

NVIDIA

Rumored acquirer of Reflection at $30B after buying Poolside for $12B; maintains Nemotron; pledged $2B more for European AI. O'Driscoll: "When you got $5 trillion, if you want to buy something for $30 billion, you can buy something for $30 billion." 00:21:31 Lemkin: "Salesforce announced they've got their own models that are built on Nemotron and built with NVIDIA." 00:23:06

Reflection

US open-model startup backed by Sequoia, last round about $25B, rumored NVIDIA target at $30B. Harry: "a fantastic team building open-source models in the US." 00:20:01 O'Driscoll said he'd "hit the bid." 00:21:57

Poolside

Open-model company NVIDIA bought for $12B. Lemkin referenced its sale rationale: "The sort of Poolside, 'Why We're Selling' memo that we just couldn't get enough compute, we just couldn't afford it, we could not keep up with the arms race." 00:24:50

Thinking Machines

Raising at $44B. Lemkin questioned compute access: "Can you even get the compute to be competitive?... You can't even get memory, let alone compute." 00:24:50 O'Driscoll noted it "has grown nicely in revenue." 00:23:52

Jev (and its parent Typeface)

Raised an $870M Series A at $7.5B, led by Andreessen with Sequoia participating; a fast, cheap classifier-style product. Lemkin: "Jev usage has quadrupled just since Andreessen actually shook hands on the deal." 00:25:55 He described it: "Jev is super fast and super cheap, okay? And it is accurate enough to use for maybe 10% of your AI use cases." 00:29:08 O'Driscoll's math: "that's a $4 to $5 billion revenue company, which is probably worth 50 billion." 00:29:45

Andreessen Horowitz

Led the Jev round; credited for a winners-only math. Lemkin: "People mocked Andreessen in the early days for overpaying... 'Yes, you have to be in all the winners... the super winners will make up for all of it.'" 00:35:52

Cloudflare

Competitor to Jev, and a case study in a well-executed layoff. Lemkin: "Matthew just came in and said, 'I'm laying off 20% of the team'... They've accelerated growth even more. The stock price is up like 80% this year." 00:56:15

Amazon / OpenAI / Anthropic / OpenRouter

Named as potential Jev competitors. Lemkin: "Are we worried that OpenAI appears to be having a competitive product?... Amazon has suggested they're building their own version of Jev?" 00:25:55

Isomorphic Labs (DeepMind)

Demis Hassabis's AI drug-discovery company, in talks to raise at $40-50B after about $10B from Thrive. O'Driscoll: "here the value capture is not replace labor. It's actually make science discoveries." 00:37:26

Thrive Capital

Backed Isomorphic's prior raise. Harry: "raised it about $10 billion, give or take, from Thrive matter of months ago." 00:36:19

Mecha

Physical-world/robotics data provider, a $60M Series B from Sequoia, over $100M ARR. Harry: "Mecha, which is the Mercer or the Serge for robotics, physical world data." 00:40:29

MicroAGI

European robotics data company. Harry: "MicroAGI in Europe, which has raised to $100 million very quickly." 00:40:29

Mercor / Surge

Cited as data-vendor analogues. 00:40:29

Generalist, Pi, Skild

Robotics foundation-model companies. O'Driscoll: "if Generalist, Pi, and Skild become the dominant foundation models for robotics, then it is equivalent to my foundation model comment." 00:45:11

Figure / BMW / GXO

Humanoid deployment datapoints. O'Driscoll: "'42 POCs of humanoids, zero deployed.' And then there was some thread about the BMW relationship with Figure where, again, we talked a big game, but the humanoids aren't out there." 00:42:15

Anthropic (Haiku 5.5 / Sonnet 5.5 / Opus)

Cut Haiku prices by 90% in the OpenAI price war. Lemkin: "Haiku 55 is almost as good for a lot of use cases as Sonnet and it's a 10th or the 20th the price." 00:46:44 Also noted: "Anthropic's reaching out to Muse, seeing Meta, 'Can we get some of your computes?'" 00:24:50 And "we never really got the full story on Anthropic's wet lab." 00:38:34

Salesforce

Building models on Nemotron and moved against open-weight Chinese dependencies. Lemkin: "Salesforce making Finn rip it out before the deal closed is telling." 00:51:35

Bending Spoons

Acquirer known for deep layoffs (Airtable, Vimeo, Evernote, Miro mentioned). Lemkin: "you know what the Bending guys are going to do. They're going to just axe almost your entire team." 00:55:11 Harry noted the strategic constraint: "it actually has to find large enough companies that it can buy." 00:56:58

Airtable, Miro, Vimeo, Evernote

Companies referenced as Bending Spoons acquisitions or targets. 00:52:25

Constellation Software / Vista / Thoma Bravo / Francisco Partners

Named as alternative acquirers for comparison. 00:55:11

Bumble

Dating app being taken private by Ravi Gupta's Ithaka Holdings for about $1.3B. Lemkin's AI bull case: "if you can get other data and you can use AI to make dating better... Raise prices because it's much better." 00:59:02

Ithaka Holdings

Ravi Gupta's new firm acquiring Bumble. 00:57:45

Flock Safety

Surveillance company cutting 270 people (18%) amid privacy backlash. Lemkin cited security failings: "stored images unencrypted, the cameras sent credentials in clear text... the police interface didn't require any authentication." 01:08:21

TechWolf

Belgian HR-AI company that sold to SAP for hundreds of millions, a 20VC portfolio company. Harry: "it was able to transition into an AI-first company and sold to SAP... the biggest ever for Belgium." 01:03:30

SAP

Acquirer of TechWolf. 01:03:30

Intercom (and Finn)

AI-native reboot to nine figures in revenue. Lemkin: "Kudos to the Intercom for building an AI-native reboot of the company to nine figures in revenue." 01:04:14

Menlo Ventures

Cited for fund math. Lemkin: "they're betting they'll have at least four $25 billion plus exits to make the fund." 00:34:34

Coinbase

Referenced as the investment that powered Garry Tan's IRR via pro rata. 00:10:37

Quinn Emanuel

Law firm retained by Day One; "notorious firm." 00:05:49

Zero Hash, Deel, Framer

Episode sponsors. Zero Hash powers Kalshi, Stripe, Morgan Stanley, Gusto, Interactive Brokers; Deel IT; Framer used by Perplexity and Miro. 00:01:03

Nasdaq / World Cloud / saastr.ai index

Used for SaaS recovery context: "Nasdaq's up 22% this year... World Cloud's up 25%." 01:01:10

4. People Identified

Founder of SaaStr and co-host. Mentioned as a builder-investor; also shared a personal pro rata story. His tactical edge: heavy daily use of AI tools informs his takes. "I'm using Jev a lot, and I can already see that if it never changed, maybe I would swap it out." 00:28:28

Rory O'Driscoll

Co-host (Scale Venture Partners). Supplied the decacorn data and pricing logic: "There are in the public markets a grand total of 98 decacorns ever." 00:31:41

Harry Stebbings

Host and founder of 20VC. Brought the TechWolf exit and the news agenda. 01:03:30

Keith Rabois

Investor; Lemkin credited his pro rata framing as the only thoughtful take: "I'd have to pay like half the price I'm paying otherwise." 00:15:16

Ali Partovi

Seed investor (Neo) who publicly waived pro rata rights. Harry: "a fantastic investor, who's made a lot of money for his investors." Lemkin: "I nominated him last year for Investor of the Year... his fund doesn't need them!" 00:09:21

Garry Tan

YC CEO; his comment that pro rata rights made his IRR on Coinbase was noted by O'Driscoll. 00:10:37

Masha

Day One Ventures investor; Lemkin's co-investor in Owner. "She's a good investor." 00:05:49

Mark Zuckerberg

Meta CEO. Lemkin: "I think Zuck's had a lonely stretch... it's even harder to find someone you can truly trust." 00:00:00

Alex Wang

Scale AI founder, now at Meta. Lemkin: "It's not easy to pay someone $10 billion and have him stick around and work that hard." 00:18:43

Sheryl Sandberg

Former Meta COO; referenced as Zuckerberg's earlier right-hand. 00:19:30

Nat Friedman / Daniel Gross

Meta AI leaders mentioned by Lemkin. 00:19:30

Jensen Huang

NVIDIA CEO. His $2B European AI commitment: "anything you can spend to deflect any anti-Americanism is probably worth the money." (Lemkin) 01:09:42

Demis Hassabis

DeepMind/Isomorphic founder. Lemkin: "generational talent has generational prices in today... he's going to pull together a team like's never been seen since the dawn of mankind in artificial intelligence." 00:36:40

Marc Benioff

Salesforce CEO; made the point at the SaaSpocalypse low that "this too will pass." 01:01:10

Matthew Prince

Cloudflare CEO; executed a 20% layoff while accelerating growth. 00:56:30

Ravi Gupta

Founder of Ithaka Holdings, acquiring Bumble. 00:57:45

Bryan Johnson

Longevity influencer; cited as the reason Lemkin was skeptical of longevity medicine. 00:39:21

5. Operating Insights

Don't just evaluate on benchmark parity; evaluate whether the vendor will keep investing

Lemkin's Jev analysis offers a general rule for competing with giants: parity is easy to hit once, hard to sustain. "It's easy to build what OpenRouter was six months ago today, right? But are they going to maintain that investment?" 00:28:28 For operators, bet on the team that does nothing else.

Run a one-time, decisive layoff from strength, not as a slow bleed

Lemkin's Cloudflare case: the CEO cut 20% "not because we're struggling, but because we're [trying] to grow even faster," and they "beat the quarter" and "accelerated growth even more." 00:56:30 Smaller 5-7% cuts can be too little to change the operating model.

Negotiate employee economics upfront if you sell

O'Driscoll's tactical advice for a founder facing a layoff-heavy acquirer: "cut your deal on your employee economics upfront, so know what everyone's getting into." 00:54:26

Treat unsecured basics as existential for security-adjacent companies

Lemkin on Flock: "when you're ultimately a security company... there are certain companies where you can't cut too many corners." Board and investors should probe the corners cut, not assume buyers will. 01:08:21

Swap in the cheaper, faster model now and demand COGS results

Lemkin to AI-product operators: with Haiku and Sonnet pricing and speed, "If it's twice as fast and, in some cases, a 10th the price, we sure better see a lot of results at the next board meeting. No more excuses." 00:48:19 Pair with O'Driscoll's elasticity test: a 90% cut requires a 10x volume increase to be revenue-neutral. 00:50:06

6. Overlooked Insights

The "strategic investor" spring-back right is the real mechanism that broke the Valar round

Buried in the discussion is that the dispute wasn't about pro rata per se but about founder favoritism. Lemkin: "if they're strategic, they're already on the cap table. You got the strategy! ... This is the whole point, right? This is, you're not supposed to play favorites." 00:06:42 O'Driscoll's earlier definition makes the structural risk explicit: spring-back rights exist precisely so a waiver vote can't be rigged by letting one holder buy in afterward. 00:05:36 Any founder running a "everyone waives" process who lets one insider buy in has created a legal and relational liability.

Haiku 5.5's reasoning-dial changes B2B unit economics, not just chat costs

Lemkin noted in passing that Haiku 5.5 is "almost the first reasoning model, and you can dial up how much reasoning you want, that you can almost not care about what it costs." 00:47:08 The overlooked implication is operational: workflows previously sampled for cost reasons ("you'll use it in every row of the database," per O'Driscoll's colleague) can now run exhaustively, which shifts product design from sampling to full coverage and favors whoever redesigns workflows first. 00:50:23