Synthesis

Frontier-lab valuations

The hosts' debates on OpenAI's and Anthropic's valuations, fundraising and IPO prospects, and what would have to be true to justify them.

Type
Synthesis
Confidence
Low
Sources
21 videos
First published
Updated

The largest cluster here: 21 episodes and 42 speakers, though the argument is largely the hosts’. Sam Lessin says pricing the labs financially is impossible because leads are easy to lose and projection variance is huge, so valuations are high-variance narrative ; on a trillion-dollar listing, ‘all you have to believe is that number go up’ . Reading a chart in which Anthropic’s EBITDA reaches 48 billion in 2030, he says it would not justify the valuation . He judges revenue by whether the slope is linear or super-linear . Jess Lessin reports valuation rumours (Anthropic at 800 billion) and says the likely outcome is a valuation reset . On government stakes, Sam reads an OpenAI offer of 5 percent as a distress signal, which Dave puts down to a press headline (C8). Dave warns of the risk riding on frontier-lab positions . Treat all valuation numbers as spoken, often currency-unstated or from memory , and note Jess says she is long both labs.

Key points

  • Sam says valuations cannot be priced financially and rest on narrative and cheap money — 4 episodes. (Sam · Stripe Buys OpenRouter, OpenAI (Aug 2026) ▶ 30:48)
  • Judge growth by the slope of revenue, and check how ARR is calculated — 3 episodes. (Sam · Anthropic’s Bet on Coding Is W (Mar 2026) ▶ 12:20)
  • Raises and IPOs are discussed at ever larger rumoured figures — 6 episodes. (Jess · Allbirds Pivoted to AI Data Ce (Apr 2026) ▶ 25:22)
  • Burn and revenue figures for OpenAI are cited as evidence of the strain — 3 episodes. (Jess · SpaceX Hit $2.5 Trillion. Now (Jun 2026) ▶ 21:45)
  • A government stake is read as distress by Sam and as PR by Dave — 1 episode. (Sam · AI Frontier Lab Is Essentially (Jul 2026) ▶ 19:41)
  • Some think the labs will not fall to zero and a reset is likelier than a collapse — 3 episodes. (Sam · OpenAI vs Google vs Meta: Busi (Nov 2025) ▶ 01:04:19)

Where guests disagree

Is the OpenAI stake offer a distress signal? ⚠️

  • Governments only take stakes in distressed companies, so it signals distress (Sam) (Sam · AI Frontier Lab Is Essentially (Jul 2026) ▶ 19:41)
  • It is a press headline that everyone gets a piece of AI (Dave); Jess says the headline is directionally right but specifically wrong (Dave · AI Frontier Lab Is Essentially (Jul 2026) ▶ 09:56)

Frameworks

  • Linear versus super-linear slope — Sam’s investor lens on AI revenue charts: the level matters less than whether the curve is linear or asymptoting (bad) or exponential (maybe good). (Sam · Anthropic’s Bet on Coding Is W (Mar 2026) ▶ 12:20)
    • Anthropic’s recent months look like a super-linear curve with no asymptote
    • OpenAI’s chart looks like a company scraping to make the number go up
    • The chart to watch is what it looks like in two months
  • Labs as stealth bombers: zero-day models held in reserve — Sam’s thesis for how the model labs become valuable over 5 to 10 years. (Sam · Anthropic Wants to Slow Down. (Sep 2026) ▶ 18:52)
    • They cannot make money competing in the token marketplace
    • With about 10,000 PhDs they develop zero-day models against other LLMs and never release them
    • Like nuclear weapons: when a country is subverted, only one or two parties can fix it, and they can charge huge sums
    • Zero-day only works while the model is unexposed, as exposed models get built on

Stories & examples

  • Government stakes in distressed companies: Sam recalls the US taking stakes in banks in 2008, airlines, car companies and Intel to bail them out because they were important. Lesson: The only time governments do this is when firms are distressed, so OpenAI offering equity without cash reads as distress. (Sam · AI Frontier Lab Is Essentially (Jul 2026) ▶ 19:21)

Numbers

  • Anthropic valuation at the chart’s investment point: 360 billion (currency unstated) (host’s framing, currency unstated) (Dave · OpenAI’s $852 Billion Valuatio (Apr 2026) ▶ 35:59)
  • Anthropic EBITDA 2030 (chart): 48 billion per year (host’s framing) (Dave · OpenAI’s $852 Billion Valuatio (Apr 2026) ▶ 36:16)
  • OpenAI funding round being closed: 122 billion dollars (host’s framing) (Jess · OpenAI’s $852 Billion Valuatio (Apr 2026) ▶ 07:23)
  • Anthropic raise valuation (Jess heard): 800 billion (currency unstated) (host’s framing, approx., currency unstated) (Jess · Allbirds Pivoted to AI Data Ce (Apr 2026) ▶ 25:22)
  • OpenAI cash burn, Q1: 4 billion (currency unstated) (host’s framing, currency unstated) (Jess · SpaceX Hit $2.5 Trillion. Now (Jun 2026) ▶ 21:45)
  • OpenAI revenue, Q1: 5.7 billion (currency unstated) (host’s framing, currency unstated) (Jess · SpaceX Hit $2.5 Trillion. Now (Jun 2026) ▶ 21:47)

Caveats & warnings

  • Jess is unsure of the Anthropic valuation she cites and says she should look it up. (Jess · OpenClaw vs Meta vs OpenAI: Th (Feb 2026) ▶ 23:01)
  • Jess warns that ARR can be calculated many ways; the chart figures are generous and on a 4-week basis, though she thinks the two companies are apples to apples. (Jess · Anthropic’s Bet on Coding Is W (Mar 2026) ▶ 13:25)
  • Sam says he has no idea what price he would pay for Anthropic stock with a five-year lockup, and that no one does. (Sam · Stripe Buys OpenRouter, OpenAI (Aug 2026) ▶ 31:53)
  • Jess says her view on Anthropic, OpenAI and SpaceX is as of today and she reserves the right to change her opinion tomorrow. (Jess · Google I/O Just Killed SEO? Ag (May 2026) ▶ 41:54)
  • Hosts say they do not provide investment advice on the pod. (one of the hosts · Google I/O Just Killed SEO? Ag (May 2026) ▶ 33:35)
  • Sam says he has no opinion on OpenAI’s valuation beyond that it is fine and money is cheap. (Sam · OpenAI’s $852 Billion Valuatio (Apr 2026) ▶ 33:48)

Also covered here: Core business plus option value

Five episodes, with the formal statement in a single one (fAb3YAwHYPI). Sam says public AI companies are bought for a core business plus option value (F3). His test of a supplier’s moat is how much you would miss it: if Claude vanished tomorrow he would care for about a second (P5, L2), even though he uses it as the core of his infrastructure . He says Anthropic’s core premium-token business has been flatlining for months , a point another host admits is only Twitter hearsay . Earlier, Sam argued that having lots of profit gives many shots on goal, which OpenAI lacks , and Jess Lessin predicted big tech and AI startups would diverge in 2026 . An unnamed host says mega-companies need enough business lines to justify valuations . Treat this as one host’s framing rather than a shared model.

Key points

  • Sam values large AI companies as a core business plus option value — 2 episodes. (Sam · Adrian (Oct 2026) ▶ 18:57)
  • Test a supplier’s core by how much you would miss it — 1 episode. (Sam · Adrian (Oct 2026) ▶ 21:27)
  • Profit buys many shots on goal that a pure lab lacks — 2 episodes. (Sam · Google Just Won AI, TikTok’s “ (Nov 2025) ▶ 27:38)
  • Anthropic’s core is reported as flat, but only second-hand — 1 episode. (Sam · Adrian (Oct 2026) ▶ 19:21)

Frameworks

  • Core business plus option value — Sam’s model for judging large public AI companies: a core business you can price, plus option value from several things that might make it valuable. (Sam · Adrian (Oct 2026) ▶ 18:57)
    • Core business: selling premium tokens
    • Option value: six things that might make the company super valuable
  • AI tide lifted all big tech; fortunes diverge in 2026 — Jess’s one framework for her 2026 predictions: the last 18 months of AI lifted big tech together, and next year the fortunes diverge dramatically. (Jess · Code Red, Bitcoin Costs, AWS R (Dec 2025) ▶ 26:26)
    • Last 18 months: AI lifted tech companies in a similar wave, hype and fortunes
    • Press coverage made companies look evenly matched in head-to-head battles
    • 2026: results, investments and strategies diverge; the tide goes out for some and keeps rising for others

Caveats & warnings

  • The speaker only repeats what they saw on Twitter about Anthropic’s flatlining revenue. (one of the hosts · Adrian (Oct 2026) ▶ 06:36)

Also covered here: Why labs cannot slow down

Five episodes, essentially Dave Morin’s thesis with Jess Lessin and Sam Lessin adding. Dave says capital wants a return, so labs cannot slow down even when they say they should (QT4), and that the capital in Anthropic over the last 12 months will be underwater if growth drops . Jess says Altman keeps saying the pace must slow but nothing follows . Dave compares it to ZIRP-era SaaS startups and describes recursive self-improvement as another reason to keep going (D1). Another Dave point is that open source lags the frontier, so a frontier lab must keep making the next model . Sam links lab behaviour to the West’s discomfort with zero and infinity and says the race-to-AGI story was told to raise money . Thin evidence: one anchor episode, and the Altman remark is Jess’s reading rather than a quote.

Key points

  • Dave argues capital wants a return, so labs race even when leaders say they should slow — 1 episode. (Dave · iPhone Duo, Meta Muse & Why AI (Sep 2026) ▶ 37:37)
  • Dave links it to prior capital cycles (ZIRP-era SaaS) and to self-improvement hopes — 1 episode. (Dave · Jeff Dean Leaves Google, Airta (Aug 2026) ▶ 30:32)
  • Open source trailing the frontier forces labs to keep releasing the next model — 1 episode. (Dave · Elon Musk vs OpenAI Trial, Goo (May 2026) ▶ 30:23)
  • Sam says the AGI-race story was a fundraising story — 2 episodes. (Sam · OpenClaw vs Meta vs OpenAI: Th (Feb 2026) ▶ 46:51)

How much to trust this

  • Evidence: 26 episodes, 2 independent speakers (the four hosts and unidentified host speech count as one voice, the show; guests counted individually).

Sources

  1. (01)

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

    YouTube

    26:26 into “Code Red, Bitcoin Costs, AWS Resurgence, and Waymo Safety (Not About AI... ish)”

  2. (02)

    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”

  3. (03)

    Stripe Is Trying To Buy PayPal For $150b | OpenAI Cyber Incident, Frontier Model Economics, Peptides

    YouTube

    31:44 into “Stripe Is Trying To Buy PayPal For $150b | OpenAI Cyber Incident, Frontier Model Economics, Peptides”

  4. (04)

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

    YouTube

    29:34 into “iPhone Duo, Meta Muse & Why AI Finally Scared Dave Morin”

  5. (05)

    OpenAI vs Google vs Meta: Business Model War

    YouTube

    57:37 into “OpenAI vs Google vs Meta: Business Model War”

  6. (06)

    Elon Musk vs OpenAI Trial, Google Cloud Surge, Meta’s Blocked Acquisition, Anthropic Winning

    YouTube

    13:29 into “Elon Musk vs OpenAI Trial, Google Cloud Surge, Meta’s Blocked Acquisition, Anthropic Winning”

  7. (07)

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

    YouTube

    25:22 into “Allbirds Pivoted to AI Data Centers | SpaceX $1.5T IPO, Anthropic $800B, $175M Seed Rounds”

  8. (08)

    OpenAI's $852 Billion Valuation Makes Zero Sense (ClawCon Tokyo, Anthropic's Leak, Goodbye Internet)

    YouTube

    7:23 into “OpenAI's $852 Billion Valuation Makes Zero Sense (ClawCon Tokyo, Anthropic's Leak, Goodbye Internet)”

  9. (09)

    The Internet Is Fighting Back Against Your AI Agent

    YouTube

    6:50 into “The Internet Is Fighting Back Against Your AI Agent”

  10. (10)

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

    YouTube

    30:48 into “Stripe Buys OpenRouter, OpenAI vs. Anthropic & the AI Backlash”

  11. (11)

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

    YouTube

    21:45 into “SpaceX Hit $2.5 Trillion. Now Watch What Elon Buys Next.”

  12. (12)

    ⁠The Real Reason Elon Wants SpaceX and Tesla to Merge Now

    YouTube

    21:18 into “⁠The Real Reason Elon Wants SpaceX and Tesla to Merge Now”

  13. (13)

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

    YouTube

    10:50 into “Jeff Dean Leaves Google, Airtable Sells, and the Bag-Securing Era”

  14. (14)

    Google Just Won AI, TikTok’s “Self-Control Badges”, ChatGPT Enters Group Chats

    YouTube

    25:57 into “Google Just Won AI, TikTok’s “Self-Control Badges”, ChatGPT Enters Group Chats”

  15. (15)

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

    YouTube

    23:17 into “Google I/O Just Killed SEO? Agentic Search, Universal Cart, and the New Web”

  16. (16)

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

    YouTube

    36:53 into “Chinese AI Model GLM 5.2 Beating Frontier Models | Meta Glasses, Polymarket Scandal, AI Talent War”

  17. (17)

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

    YouTube

    6:36 into “Meta Muse vs Instinct vs OpenAI Dots: Do AI Agents Have Any Moat?”

  18. (18)

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

    YouTube

    12:30 into “Anthropic Wants to Slow Down. Is It Because They're Losing?”

  19. (19)

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

    YouTube

    12:20 into “Anthropic’s Bet on Coding Is Working (OpenAI Shopping Pivot, A16Z’s Top 50 List, $1B Tennis Channel)”

  20. (20)

    Silicon Valley Is Great at Tech and Terrible at Business

    YouTube

    14:02 into “Silicon Valley Is Great at Tech and Terrible at Business”

  21. (21)

    AI Frontier Lab Is Essentially Doomed | Open AI Offered U.S. Equity, Xbox Is In Trouble, Post-AGI

    YouTube

    5:22 into “AI Frontier Lab Is Essentially Doomed | Open AI Offered U.S. Equity, Xbox Is In Trouble, Post-AGI”

(By email)

Compounding Insights