Anthropic IPO at Risk, Meta’s Muse Pop, Token Prices Fall, Open Source Gains Share, Alignment Fails

All-In Podcast · 2026-09-26 · 94 min
https://www.youtube.com/watch?v=cvP_1jmnkmMVideo summary
Anthropic IPO faces liquidity pressure as open-source AI flips token usage 80/20, while Meta Muse reaches 3 million downloads.
The panel argues that Anthropic and OpenAI are no longer merely “labs” but corporations that must accept product liability and investor scrutiny. Anthropic’s planned IPO is clouded by a potential $2 trillion valuation, warnings of more than 10% extinction risk, a San Francisco wet lab, and debate over founder super-voting shares. Open-source competition is accelerating: token usage reportedly flipped from 80% closed to 80% open in 12 weeks, while models such as DeepSeek, Qwen, Mimo, and Bonsai 2 approach premium performance at radically lower cost. The hosts say frontier firms must move beyond selling tokens into cybersecurity, law, and specialized science. They also reject Bernie Sanders’ proposed super-intelligence ban as unenforceable and economically damaging. Meta’s Muse offers a consumer counterpoint, reaching number one in the App Store and 3 million downloads in about 10 days, with practical agent features such as email triage, travel booking, shopping, and subscription management. Anthropic’s wet lab is described as a low-risk BSL1/BSL2 facility using AI to discover and test enzymes, not create pathogens.
Chapters
- 0:00Bestie Intros: Trump’s spontaneous Jensen Huang call at the All-In Summit
- 3:54AI liability: Frontier companies must stop calling themselves labs
- 9:15AI liability lobbying: Companies seek Section 230-style protection
- 15:03AI competition and the Wuhan lab comparison
- 20:03Open-source AI surge: DeepSeek, Qwen, Mimo, and Muse arrive in ten days
- 24:45AI commoditization: Model convergence forces OpenAI and Anthropic up the stack
- 29:22Anthropic and OpenAI postpone IPOs: Anthropic’s 2026 odds fall to 76%
- 34:09Anthropic and OpenAI postpone IPOs: IPO pricing may fall below $2 trillion
- 39:03Anthropic and OpenAI postpone IPOs: open-source tokens flip to 80% share
- 44:01Anthropic and OpenAI postpone IPOs: frontier models face a six-month hamster-wheel risk
- 48:36Anthropic and OpenAI postpone IPOs: open source and costly tokens pressure margins
- 53:11Political reactions to “Pacing the Frontier”: Bernie Sanders proposes a superintelligence ban
- 57:14Political reactions: Obama’s $10 trillion frontier-company wealth thesis
- 1:02:59Political reactions: Bernie Sanders’s AI ban compared with China’s shipbuilding ban
- 1:07:58Meta launches Muse: three million downloads and 10% Meta stock gain
- 1:13:03Meta launches Muse: agents challenge Amazon, App Store fees, and subscription friction
- 1:17:30Meta launches Muse: agent adoption, alignment research, and Oracle’s data-center warning
- 1:22:49Meta launches Muse: Anthropic’s Claude constitution and the AI investment cycle
- 1:27:13Anthropic’s bio research and “wet lab” in SF: AI-discovered CRISPR-like enzyme tested at BSL1/BSL2
This is a Tier 1 public summary
Whether the chapter key points, section summaries and mind map are public is up to the person who shared it. Want the full analysis?Submit one yourself.
More from this channel
Jensen Huang: The Doomer Hoax, Superintelligence is Here, and The Future of AI (ft. President Trump)All-In PodcastJensen Huang rejects AI doom predictions, backs open models, and says superintelligence is already here in narrow domains.
Elon Musk & Gwynne Shotwell on AI Risks and Peer Review, Starship, Terafab, SpaceX/Tesla MergerAll-In PodcastElon Musk forecasts Mars landings within a decade while SpaceX targets Starship reuse in 2027 and AI peer review
JD Vance on AI, Entitlement Fraud, Iran War, Israel, H-1B Abuse & the MidtermsAll-In PodcastJD Vance touts 3 million fewer border crossings, $250 billion saved, and a $100,000 H-1B fee.
Satya Nadella on the AI Doomer Slowdown, Microsoft’s Master Plan & Who Wins AIAll-In PodcastSatya Nadella outlines Microsoft’s $175 billion AI buildout, 30 million Copilot users, and safer multi-model infrastructure
Meta's Dina Powell McCormick: The Case for Data Centers, Backlash, AI Job Boom & Meta’s FutureAll-In PodcastDina Powell McCormick defends Meta’s data centers, $50,000 teacher bonuses, AI jobs, teen limits, and Meta glasses
Related analyses
EP313. 谷歌痛失 AI 大將、Fable 5 配方解密、Meta 士氣低落中、FSD 即將入台 | M觀點M觀點Google痛失AI大將,Fable 5靠Agent Harness變強,Meta士氣崩盤、特斯拉FSD送審台灣
EP336. 前沿模型該減速嗎、Anthropic 威脅報告、為什麼要一直聊模型 | M觀點M觀點Anthropic倡議前沿AI控速,Kimi疑將解放軍影像轉交Claude,AI硬體需求反而可能增加
EP337. 微軟的 AI 憲法、PLTR 輝達合作、逆風看法的本質 | M觀點M觀點微軟發布AI憲法,Palantir攜NVIDIA用開源模型挑戰OpenAI與Anthropic
2026/8/19(三)債市風暴 撿便宜時刻?美中爭霸燒錢戰 Al時代只能二選一?【早晨財經速解讀】游庭皓的財經皓角美債殖利率飆升,AI融資、伊朗戰事與中國模型掀起科技股估值回檔
To Regulate AI Effectively, Focus on How It’s Useda16z Deep DivesMartin Casado argues AI laws should target harmful use, while US regulatory uncertainty pushes startups toward Chinese open-source models.