Blueprint

Seed investing in an AI market

What the hosts and a few guest investors say about seed pricing, fund maths, judging founders and exits in the AI boom, with each figure labelled as reported and each view tied to a fund stake.

Type
Blueprint
Confidence
Medium
Sources
26 videos
First published
Updated

Voice and stakes

Voice: mixed (Sam, ~39 % of attributed items; Brit, ~19 %). The argument runs mainly between two people who run funds, with Dave, Jess and a handful of guest investors adding cases.

Stakes. Sam’s firm is Slow Ventures, and he says Slow was an early investor in Airtable and an Allbirds investor that sold earlier (Sam · Jeff Dean Leaves Google, Airta (Aug 2026) ▶ 21:50) (Sam · Allbirds Pivoted to AI Data Ce (Apr 2026) ▶ 09:08). Brit’s fund is Offline, which she says will raise its next fund soon (Brit · Allbirds Pivoted to AI Data Ce (Apr 2026) ▶ 38:00). A host notes that Groq is an Offline company (one of the hosts · Founders Beware: Illiquid Valu (Jan 2026) ▶ 14:34). Guest Peter Deng’s firm, Felicis, was first cheque into Periodic Labs (Peter Deng · Peter Deng & Rachel Holt (Aug 2026) ▶ 43:05). Nobody here is a neutral observer of seed pricing.

The argument in brief

  1. Brit cites Carta that the top 5 % of seed rounds sit above a $175 million valuation, three times last year and double 2021 (Apr 2026, as she states it) (Brit · Allbirds Pivoted to AI Data Ce (Apr 2026) ▶ 36:31). She calls it the bubble of all bubbles (Brit · Allbirds Pivoted to AI Data Ce (Apr 2026) ▶ 36:17).
  2. Sam says he does not buy that a $480 million seed raise at $4.48 billion makes money, because Slow’s big wins were all under $10 million pre-money (Jan 2026) (Sam · The Seed Market Is Broken: Meg (Jan 2026) ▶ 06:21).
  3. Slow’s lane is $1-3 million for about 10 percent, which Brit calls the fairway of seed (Sam · The Seed Market Is Broken: Meg (Jan 2026) ▶ 07:47).
  4. The top AI outcomes are said not to return a seed fund: Sam puts OpenAI seed investors at about 25x (Sam · The Seed Market Is Broken: Meg (Jan 2026) ▶ 08:44), and Dave says a 2023 core Anthropic cheque would not have returned a seed fund either, though roughly 80 million is still good (Dave · The Seed Market Is Broken: Meg (Jan 2026) ▶ 10:05).
  5. Dave describes a “tale of two cities”: name-brand founders raise at billion-plus valuations while a separate wave of seed and incubator activity continues (Dave · OpenClaw vs Meta vs OpenAI: Th (Feb 2026) ▶ 24:04).
  6. Brit says mainstream seed is near-perfectly priced (“synthetic beta”) because about 2,000 seed funds chase the same deals, so alpha sits in illegible spaces (Brit · SaaS Companies Beware: AI Is T (Jan 2026) ▶ 41:24) (Brit · SaaS Companies Beware: AI Is T (Jan 2026) ▶ 41:35).
  7. Sam says he is “currently wrong” because the Valley is trading, not investing, and he bets he becomes right (Sep 2026) (Sam · Anthropic Wants to Slow Down. (Sep 2026) ▶ 31:19).

Pricing: what the hosts say is happening

The figures in this section are reported or second-hand. Brit relays the Carta statistic and says top-decile Series B to D valuations are 200 to 300 % above 2021 (Brit · Allbirds Pivoted to AI Data Ce (Apr 2026) ▶ 37:24). In the January discussion of the $480 million seed, the valuation and round size are taken as given (Brit · The Seed Market Is Broken: Meg (Jan 2026) ▶ 05:37) (Brit · The Seed Market Is Broken: Meg (Jan 2026) ▶ 05:40), and Brit adds a sequence for Thinking Machines of a 2 billion seed, then 10 billion, then a raise at 50 billion (Brit · The Seed Market Is Broken: Meg (Jan 2026) ▶ 13:16). Jess relays that Anthropic was around 380 billion (she is unsure) and that Anduril was in talks at “60” (Feb 2026) (Jess · OpenClaw vs Meta vs OpenAI: Th (Feb 2026) ▶ 23:01). Someone at an episode in September says Hermes raised about 1.5 billion and agent companies are valued at 1 to 5 billion (one of the hosts · iPhone Duo, Meta Muse & Why AI (Sep 2026) ▶ 21:52).

Sam argues that late-stage funds behind the recent rounds are doing “five seconds of analysis” and are focused on access (May 2026) (Sam · Google I/O Just Killed SEO? Ag (May 2026) ▶ 24:20). He says Anthropic raising at $350 billion makes seed outcomes more top-heavy (Sam · Founders Beware: Illiquid Valu (Jan 2026) ▶ 23:37). In August he says it is the same investors and the same cap table as the SaaS boom (Sam · Jeff Dean Leaves Google, Airta (Aug 2026) ▶ 31:36); Dave likens the lab boom to the ZIRP-era SaaS loop (Dave · Jeff Dean Leaves Google, Airta (Aug 2026) ▶ 30:32). Sam calls the labs “the biggest seed companies in history” (Sam · Stripe Buys OpenRouter, OpenAI (Aug 2026) ▶ 37:30).

From the fund side, a speaker at the Upfront Summit (not identified in the corpus; Feb 2026) says there were few LPs (one of the hosts · Apple, OpenAI & Why Nobody Mak (Feb 2026) ▶ 08:02), that this is the worst fundraising market ever (one of the hosts · Apple, OpenAI & Why Nobody Mak (Feb 2026) ▶ 08:45) and that the seed market is binary, either a 50 million pre or very cheap (one of the hosts · Apple, OpenAI & Why Nobody Mak (Feb 2026) ▶ 12:03). The same speaker is unsure one anecdote from the summit is true (one of the hosts · Apple, OpenAI & Why Nobody Mak (Feb 2026) ▶ 08:39).

Fund maths and pro rata

Brit says portfolio construction is breaking under inflated valuations, so she measures the percentage of the fund per bet rather than ownership, and LPs and GPs argue over marks (Brit · Allbirds Pivoted to AI Data Ce (Apr 2026) ▶ 40:03) (Brit · Allbirds Pivoted to AI Data Ce (Apr 2026) ▶ 39:42). She also says companies that command higher valuations are statistically the most successful (Brit · Allbirds Pivoted to AI Data Ce (Apr 2026) ▶ 39:25).

Sam says pro rata rights are critical to Slow’s model and must be made explicit (Oct 2026) (Sam · Silicon Valley Is Great at Tec (Oct 2026) ▶ 42:02). His advice is to get written confirmation and tell founders you will enforce them (Sam · Silicon Valley Is Great at Tec (Oct 2026) ▶ 41:15). His example is a large VC, and an LP of Slow, that tried to waive Slow’s pro rata early in its history (Sam · Silicon Valley Is Great at Tec (Oct 2026) ▶ 41:39). Brit says big Series A funds renegotiate and the only protection is a good founder relationship (Brit · Silicon Valley Is Great at Tec (Oct 2026) ▶ 42:47). She describes Megalith as a vehicle that buys pro rata rights seed funds cannot or will not exercise, for about 1.25 % of carry (Brit · Silicon Valley Is Great at Tec (Oct 2026) ▶ 40:26) (Brit · Silicon Valley Is Great at Tec (Oct 2026) ▶ 43:30), and says about 64 % of LPs want co-invest rights, with no source given (Brit · Silicon Valley Is Great at Tec (Oct 2026) ▶ 44:22). Sam sums up private markets as “the treaty markets” (Sam · Silicon Valley Is Great at Tec (Oct 2026) ▶ 45:48).

On staying capital-light, Sam says a seed investor should avoid “super capital-consumptive” companies (Sam · The Seed Market Is Broken: Meg (Jan 2026) ▶ 08:57). Peter Deng says over 20 % of Felicis capital in recent funds went to resilience, energy, manufacturing and defence (Peter Deng · Peter Deng & Rachel Holt (Aug 2026) ▶ 49:58), and Sam expects the middle of the startup market to drop out after the SpaceX listing (Jun 2026) (Sam · SpaceX Hit $2.5 Trillion. Now (Jun 2026) ▶ 16:22). These are different fund types, not a head-on debate.

Judging founders

  • Team and theme. Sam says in early-stage investing you love team and theme (Sam · Suds (Sep 2026) ▶ 44:20), and that the leader must actually care about the business (Sam · Code Red, Bitcoin Costs, AWS R (Dec 2025) ▶ 41:14).
  • Story of infinity. He favours companies with a fundamental story of infinity over a fine money machine, because seed outcomes are more top-heavy than ever (Sam · Founders Beware: Illiquid Valu (Jan 2026) ▶ 23:15).
  • Competence and integrity. Guest investor Suds judges a founder on talent networks, customer or industry experience, honesty and expected value (Early-stage evaluation: competence, integrity and expected value (Suds)) (Suds · Suds (Sep 2026) ▶ 45:22).
  • Show working product. Jess will not engage without it: “if you can’t show me bits” (Jess · Elon’s Trillion-Dollar Move, W (Feb 2026) ▶ 42:00). Sam adds that AI has made every cold pitch look credible, so VCs struggle to screen them (Sam · Adrian (Oct 2026) ▶ 53:45).
  • Illegible places. Brit looks at bio, robotics, hardware and defence (Brit · SaaS Companies Beware: AI Is T (Jan 2026) ▶ 26:54); Guillermo Rauch looks for problems that look niche but become everyone’s (Guillermo Rauch · Guillermo Rauch (Apr 2026) ▶ 35:39).
  • Sourcing. Sam sources deals from Twitter posts and screenshot essays (Feb and Dec 2025 episodes) (Sam · Apple, OpenAI & Why Nobody Mak (Feb 2026) ▶ 12:21) (Sam · Code Red, Bitcoin Costs, AWS R (Dec 2025) ▶ 16:50); Brit says to tell people publicly what you care about (Brit · Code Red, Bitcoin Costs, AWS R (Dec 2025) ▶ 14:45).

Exits, tax and paper wealth

The Airtable sale is the one worked example. Jess puts Airtable’s peak at $11 billion and its sale to Bending Spoons at about $1.5 billion (Aug 2026, as she puts it) (Jess · Jeff Dean Leaves Google, Airta (Aug 2026) ▶ 19:47) (Jess · Jeff Dean Leaves Google, Airta (Aug 2026) ▶ 19:42). Sam says the roughly 10x write-down is brutal and that it is good to be a seed investor (Sam · Jeff Dean Leaves Google, Airta (Aug 2026) ▶ 29:52); Slow had no information rights and learned of the sale from a tweet (Sam · Jeff Dean Leaves Google, Airta (Aug 2026) ▶ 24:57). Sam admits he has not read the deal documents (Sam · Jeff Dean Leaves Google, Airta (Aug 2026) ▶ 24:25). Dave’s name for the outcome is a “split landing” (Split landing) (Dave · Jeff Dean Leaves Google, Airta (Aug 2026) ▶ 23:20), and Sam says big investors like it because it avoids a write-down (Sam · Jeff Dean Leaves Google, Airta (Aug 2026) ▶ 24:17). His lesson is to take money off the table through the rounds (Sam · Jeff Dean Leaves Google, Airta (Aug 2026) ▶ 20:44).

On Allbirds, Jess reports a fall from over 4 billion to about 40 million (currency unstated) (Jess · Allbirds Pivoted to AI Data Ce (Apr 2026) ▶ 09:55); Sam says Slow sold earlier because holding meant believing it was the next Nike (Sam · Allbirds Pivoted to AI Data Ce (Apr 2026) ▶ 09:39).

Sam’s tax argument (Apr 2026) is that selling a heavily marked-up holding costs roughly 40-50 % in tax, so you must expect it to underperform everything else by about 50 % (Sam · Allbirds Pivoted to AI Data Ce (Apr 2026) ▶ 15:48), which he argues puts a floor under prices (Tax-lock-in floor under mega-cap stocks (Sam)) (Sam · Allbirds Pivoted to AI Data Ce (Apr 2026) ▶ 17:03). He holds SpaceX (Sam · Allbirds Pivoted to AI Data Ce (Apr 2026) ▶ 16:08). He also says people who could not sell early Facebook or Anthropic stock became far richer by accident (Sam · Chinese AI Model GLM 5.2 Beati (Jun 2026) ▶ 42:49).

Cases that are not seed

Solo-founder companies appear mostly through Polsia, a Slow portfolio company (Sam · Apple, OpenAI & Why Nobody Mak (Feb 2026) ▶ 15:03). Its growth figures change by speaker: about 100K to 700K ARR in under a week (Sam, Feb 2026) (Sam · Apple, OpenAI & Why Nobody Mak (Feb 2026) ▶ 16:12), 2 million ARR in two weeks (unidentified host) (one of the hosts · Fire Sam Altman, The End of So (Mar 2026) ▶ 32:44), 4,000 active companies (Brit) (Brit · Anthropic’s Bet on Coding Is W (Mar 2026) ▶ 32:56). Brit says the hosts are “huge investors” (Brit · Anthropic’s Bet on Coding Is W (Mar 2026) ▶ 32:56). Mike Brown of Teamshares, a Slow-backed company, argues the opposite model, permanent ownership of durable businesses (Mike Brown · Mike Brown (Jun 2026) ▶ 48:59). He says he is describing a holding company, not seed (Mike Brown · Mike Brown (Jun 2026) ▶ 29:27).

How the view moved

  • Jan 2026. Sam rejects the mega-seed as a fund-returner and sets Slow’s lane at $1-3 million for about 10 % (Sam · The Seed Market Is Broken: Meg (Jan 2026) ▶ 07:47).
  • Feb 2026. Sam says Slow made more investments last quarter than in the whole previous year while other funds froze (Sam · Apple, OpenAI & Why Nobody Mak (Feb 2026) ▶ 11:36). The Upfront Summit speaker describes a weak LP market (one of the hosts · Apple, OpenAI & Why Nobody Mak (Feb 2026) ▶ 08:45).
  • Apr 2026. Brit adds the Carta figure and calls it a bubble (Brit · Allbirds Pivoted to AI Data Ce (Apr 2026) ▶ 36:31).
  • Jul 2026. Sam says “I’m too late” is a repeated mistake that stops early investments (Sam · Every CTO Is Quietly Switching (Jul 2026) ▶ 35:12).
  • Aug 2026. Rachel Holt says investors who dismissed the physical world are now crowding in and pushing up prices (Rachel Holt · Peter Deng & Rachel Holt (Aug 2026) ▶ 52:15).
  • Sep 2026. Sam says he is currently wrong but stays a seed investor (Sam · Anthropic Wants to Slow Down. (Sep 2026) ▶ 31:25).

Where the hosts disagree

  • Are application companies investable? Suds avoids the app layer and backs what extends frontier-lab revenue (Suds · Suds (Sep 2026) ▶ 32:13); an unidentified host says app companies are always seconds from being swiped by incumbents (one of the hosts · What Tim Cook’s “Executive Cha (Apr 2026) ▶ 35:35). Sam points to fast-growing boring “LLM wrappers for middle America” (Sam · Suds (Sep 2026) ▶ 48:39).
  • Infinity or durable? Sam prefers a story of infinity (Sam · Founders Beware: Illiquid Valu (Jan 2026) ▶ 23:15). Mike Brown says “we don’t like fads” (Mike Brown · Mike Brown (Jun 2026) ▶ 29:23), though he speaks as a holding-company operator.
  • Capital-hungry or not? Sam avoids it (Sam · The Seed Market Is Broken: Meg (Jan 2026) ▶ 08:57); Felicis is moving towards it (Peter Deng · Peter Deng & Rachel Holt (Aug 2026) ▶ 49:58).
  • Is a one-person company defensible? Sam says a clone could re-implement Polsia for fewer tokens so margin tends to zero (Sam · Apple, OpenAI & Why Nobody Mak (Feb 2026) ▶ 17:12); Ben Broca says Polsia is “an economy, not a tool” (Ben Broca · Henrik Werdelin & Ben Broca (Mar 2026) ▶ 18:23).

What could make this wrong

Every key number is reported, unsourced or loosely stated. Sam says he could be wrong on the OpenAI seed maths and has not run the numbers on the Anthropic case (Sam · The Seed Market Is Broken: Meg (Jan 2026) ▶ 08:37) (Sam · The Seed Market Is Broken: Meg (Jan 2026) ▶ 11:00). Dave’s 80 million has no currency (Dave · The Seed Market Is Broken: Meg (Jan 2026) ▶ 10:11). The 64 % LP figure has no source (Brit · Silicon Valley Is Great at Tec (Oct 2026) ▶ 44:22). The weak-LP-market claims come from one unidentified speaker (one of the hosts · Apple, OpenAI & Why Nobody Mak (Feb 2026) ▶ 07:49). The tax theory is Sam’s own, untested, and about positions he holds (Sam · Allbirds Pivoted to AI Data Ce (Apr 2026) ▶ 16:08). The Polsia and Teamshares material is partly self-promotion by founders the hosts back, and the Airtable case supplies most of the exits material. There is no outcome data on any “story of infinity” bet; these are practitioners’ opinions with fund interests.

How strong is the evidence

Medium to weak. The seed-pricing question (Is the seed market broken?) has only six answers from three episodes, and the Carta number is the only external statistic. Sam and Brit carry most of it, and both run funds. Independent voices are guests in one episode each: Suds, Peter Deng, Rachel Holt, Guillermo Rauch, Mike Brown. Exits rest on one case, Airtable. Fund maths rests on four episodes.

Go deeper

Sources

  1. (01)

    Is Google About to Kill Canva? │Claude Goes Rogue, Tesla vs. Waymo

    YouTube

    32:13 into “Is Google About to Kill Canva? │Claude Goes Rogue, Tesla vs. Waymo”

  2. (02)

    Code Red, Bitcoin Costs, AWS Resurgence, and Waymo Safety (Not About AI... ish)

    YouTube

    14:45 into “Code Red, Bitcoin Costs, AWS Resurgence, and Waymo Safety (Not About AI... ish)”

  3. (03)

    SaaS Companies Beware: AI Is The New UI (Anthropic's Claude Code and Cowork)

    YouTube

    26:54 into “SaaS Companies Beware: AI Is The New UI (Anthropic's Claude Code and Cowork)”

  4. (04)

    OpenClaw vs Meta vs OpenAI: The Personal Agent Wars Heat Up

    YouTube

    23:01 into “OpenClaw vs Meta vs OpenAI: The Personal Agent Wars Heat Up”

  5. (05)

    Teamshares IPO: $60M EBITDA, 92 Companies, Zero Exits

    YouTube

    29:23 into “Teamshares IPO: $60M EBITDA, 92 Companies, Zero Exits”

  6. (06)

    iPhone Duo, Meta Muse & Why AI Finally Scared Dave Morin

    YouTube

    21:52 into “iPhone Duo, Meta Muse & Why AI Finally Scared Dave Morin”

  7. (07)

    The Seed Market Is Broken: Mega-Rounds, Narrative Games, and Walkouts

    YouTube

    5:37 into “The Seed Market Is Broken: Mega-Rounds, Narrative Games, and Walkouts”

  8. (08)

    Founders Beware: Illiquid Valuations vs. Real‑World Taxes

    YouTube

    14:34 into “Founders Beware: Illiquid Valuations vs. Real‑World Taxes”

  9. (09)

    Allbirds Pivoted to AI Data Centers | SpaceX $1.5T IPO, Anthropic $800B, $175M Seed Rounds

    YouTube

    9:08 into “Allbirds Pivoted to AI Data Centers | SpaceX $1.5T IPO, Anthropic $800B, $175M Seed Rounds”

  10. (10)

    Apple, OpenAI & Why Nobody Makes Money in AI | The SaaS Apocalypse

    YouTube

    7:49 into “Apple, OpenAI & Why Nobody Makes Money in AI | The SaaS Apocalypse”

  11. (11)

    Fire Sam Altman, The End of Software Engineers, and Why AI Is All Narrative

    YouTube

    32:44 into “Fire Sam Altman, The End of Software Engineers, and Why AI Is All Narrative”

  12. (12)

    Stripe Buys OpenRouter, OpenAI vs. Anthropic & the AI Backlash

    YouTube

    37:30 into “Stripe Buys OpenRouter, OpenAI vs. Anthropic & the AI Backlash”

  13. (13)

    The Risk Of Securitizing GPUs Like Long-Lived Infrastructure & More - A MOL Special Team Edition

    YouTube

    43:05 into “The Risk Of Securitizing GPUs Like Long-Lived Infrastructure & More - A MOL Special Team Edition”

  14. (14)

    SpaceX Hit $2.5 Trillion. Now Watch What Elon Buys Next.

    YouTube

    16:22 into “SpaceX Hit $2.5 Trillion. Now Watch What Elon Buys Next.”

  15. (15)

    Jeff Dean Leaves Google, Airtable Sells, and the Bag-Securing Era

    YouTube

    19:42 into “Jeff Dean Leaves Google, Airtable Sells, and the Bag-Securing Era”

  16. (16)

    Google I/O Just Killed SEO? Agentic Search, Universal Cart, and the New Web

    YouTube

    24:20 into “Google I/O Just Killed SEO? Agentic Search, Universal Cart, and the New Web”

  17. (17)

    What Tim Cook’s “Executive Chairman” Move Really Means | Musk vs OpenAI, xAI and Cursor Rumors

    YouTube

    35:35 into “What Tim Cook’s “Executive Chairman” Move Really Means | Musk vs OpenAI, xAI and Cursor Rumors”

  18. (18)

    Chinese AI Model GLM 5.2 Beating Frontier Models | Meta Glasses, Polymarket Scandal, AI Talent War

    YouTube

    42:49 into “Chinese AI Model GLM 5.2 Beating Frontier Models | Meta Glasses, Polymarket Scandal, AI Talent War”

  19. (19)

    Vercel CEO: 70% of Our Traffic Is Now AI Agents "Nobody Was Prepared" | Anthropic, OpenClaw, OpenAI

    YouTube

    35:39 into “Vercel CEO: 70% of Our Traffic Is Now AI Agents "Nobody Was Prepared" | Anthropic, OpenClaw, OpenAI”

  20. (20)

    Meta Muse vs Instinct vs OpenAI Dots: Do AI Agents Have Any Moat?

    YouTube

    53:45 into “Meta Muse vs Instinct vs OpenAI Dots: Do AI Agents Have Any Moat?”

  21. (21)

    Anthropic Wants to Slow Down. Is It Because They're Losing?

    YouTube

    31:19 into “Anthropic Wants to Slow Down. Is It Because They're Losing?”

  22. (22)

    Anthropic’s Bet on Coding Is Working (OpenAI Shopping Pivot, A16Z’s Top 50 List, $1B Tennis Channel)

    YouTube

    32:56 into “Anthropic’s Bet on Coding Is Working (OpenAI Shopping Pivot, A16Z’s Top 50 List, $1B Tennis Channel)”

  23. (23)

    Silicon Valley Is Great at Tech and Terrible at Business

    YouTube

    40:26 into “Silicon Valley Is Great at Tech and Terrible at Business”

  24. (24)

    Elon’s Trillion-Dollar Move, WaPo Layoffs & The Super Bowl

    YouTube

    42:00 into “Elon’s Trillion-Dollar Move, WaPo Layoffs & The Super Bowl”

  25. (25)

    The New Startup Stack: One Founder + Agents | Henrik Werdelin (Audos) and Ben Broca (Polsia)

    YouTube

    18:23 into “The New Startup Stack: One Founder + Agents | Henrik Werdelin (Audos) and Ben Broca (Polsia)”

  26. (26)

    Every CTO Is Quietly Switching Off Anthropic and OpenAI | GLM 5.2, Vox Americanus, Claude Fable 5

    YouTube

    35:12 into “Every CTO Is Quietly Switching Off Anthropic and OpenAI | GLM 5.2, Vox Americanus, Claude Fable 5”

(By email)

Compounding Insights