ポッドキャスト世界の声を聴き、思考の刻みを見る。

OpenAI CFO Sarah Friar makes her All-In debut days after the company's record $122B fundraise, walking hosts Jason Calacanis, David Friedberg, and David Sacks through the IPO calculus, the Anthropic rivalry, a teased consumer hardware device from Jony Ive's team, and how OpenAI is planning compute purchases through the early 2030s. Friar's core message: an IPO is a milestone, not a destination, and the real race is building an AI intelligence layer that serves humanity at scale — with advertising eventually subsidizing access for those who can't pay. ## [00:00] OpenAI CFO Sarah Friar joins the show! Jason Calacanis opens by flagging the scale of OpenAI's March fundraise — north of $120 billion — as the largest private capital raise in history, surpassing the prior record (Saudi Aramco's ~$30B IPO) by an order of magnitude. Friar frames it succinctly: "We think AI is the biggest era that we've seen to date. We're just starting to understand what it's going to mean for global productivity." > *"Luck is whatever the preparation meets opportunity, but you got to grab it."* ## [00:31] How OpenAI thinks about its IPO timeline The question of sequencing — SpaceX going first, Anthropic filing its S-1 confidentially mid-episode — dominates the first substantive segment. Friar refuses to treat an IPO as a competitive event. Her view: an IPO is just another financing mechanism, and timing should reflect business readiness, not optics. She notes that the $122B raise was structured to give "maximum flexibility" — buying time to choose the right window rather than being forced by capital pressure. David Friedberg asks whether the private raise is the largest in history; Friar confirms it is by an orders-of-magnitude margin. Jason Calacanis then reads out Anthropic's confidential S-1 filing live. Friar doesn't blink: going first means running the SEC gauntlet first, with no guarantee of advantage. > *"No one remembers who went first, Google or Yahoo, Lyft or Uber. In the end, we're going to have to build big, sustainable, durable companies, and fundraising will be a key component of doing exactly that."* ## [03:31] OpenAI, Anthropic, Google: The AI arms race Calacanis challenges Friar directly: did OpenAI spread too thin — Sora, new gadgets, too many projects — while Anthropic moved decisively into enterprise and developer channels, seemingly pulling ahead on developer and revenue metrics? Friar pushes back on the binary framing. OpenAI's strategy is explicitly neither pure consumer nor pure enterprise: revenue is now roughly 50/50 between the two. She runs through the multi-interface thesis: ChatGPT at 900 million weekly users, Codex crossing 5 million users in a weekend from near-zero in January, Frontier for enterprise, and API access. The compounding logic is that a single underlying model serving all surfaces means more data, better personalization, declining per-token cost, and higher gross margins — all of which feed more capital back into compute access, which is itself the scarce resource that determines competitive position. > *"Our strategy is different. We are building the AI layer, the infrastructure and it's really important that there's a single foundation but then with many interfaces out into the world."* Friedberg notes that the fastest-growing user cohorts are in Africa, with Azerbaijani and Kazakh among the fastest-growing languages — an indication of how far the user base has moved beyond English-speaking tech early adopters. ## [07:43] Navigating the compute crunch and AI bottlenecks, device preview! Friedberg revisits a framing Friar coined ~18 months earlier: one gigawatt of AI compute roughly equals $10B/year of revenue for OpenAI. The question is whether that ratio still holds and what the supply picture looks like across the industry. Friar's answer: compute is scarcer than ever, the 2026 and 2027 pipeline is effectively locked, and she's already focused on securing capacity for 2030–2032. She outlines the choke points in sequence: energy and land permitting, rack and chip supply (with current memory spike as the acute bottleneck), talent pipeline, and community trust. On the latter, she details OpenAI's commitments for the Seline, Michigan 1-gigawatt data center (part of the Oracle complex, with Sam Altman cutting the ribbon that day): no rate increases for local electricity consumers, 2,500 union jobs, $1B in Michigan taxes, and $45M into Codex education credits for workers entering jobs that will require AI fluency. Friar then teases the new consumer device without naming it or showing it — designed by Jony Ive's team, targeting a substrate that replaces thumb-first smartphone interaction. "By the end of this year we will unveil it, early next year." She describes using it as feeling "very natural, but very real" in a way that's hard to articulate but immediately felt. > *"Technology is very — can be very mechanistic, but we all know great design just makes everything fade away."* The compute scarcity directly shapes product decisions: Sora's rollout was slowed because video is token-heavy and tokens were being rationed. Inference, unlike training, is increasingly meant to be global, with real-time agentic workloads requiring low-latency distributed compute. ## [15:53] OpenAI's economics Friedberg shifts to capital allocation: what is OpenAI's equivalent of the high-ROC engine that defined Amazon's warehouse flywheel or Google's search-ads loop? Friar's answer is a three-layer model — customer value creation first, then gross margin expansion via compute deflation, then capital deployment timed against the cost curve. On the deflation side, the per-token cost from GPT-4 to GPT-4.5 fell 97% in roughly two years. The latest model, o3, was priced 2x higher than its predecessor, yet customers are still seeing 20–30% cost-per-token reductions because efficiency gains outpace the price increase. The implication: today's capex decisions will be priced on tomorrow's cost profile, so underinvesting now means mispricing the outcome. Friar is explicit about how far out the planning horizon goes. Near-term (2026–27): bottoms-up revenue model with known products and pricing. Outer years: reverse the model — buy the compute, then assume revenue will follow. She cites a specific past example: a year ago she pitched investor models showing agentic revenue at $2,000/month, which nobody believed. ChatGPT Pro at $200/month was itself called impossible. Both happened. > *"The shape of the line keeps taking us by surprise to the upside."* The Michigan data center won't produce usable compute until end-of-2027 or early 2028, illustrating the 18–24 month lag between capital commitment and revenue-generating capacity. Friar's current scarcity focus is 2030–2032. ## [26:08] Push into chips, the cloud Friedberg poses the vertical integration question: as Nvidia, Google, Microsoft, and OpenAI each push into each other's layers — silicon, models, cloud, consumer — does the stack eventually merge? And does convergence make the competitive landscape simpler or more complicated? Friar frames OpenAI's evolution with a Rubik's cube metaphor. Two years ago: one cloud partner (Azure), one chip (Nvidia), one product (ChatGPT), one price point ($20/month). Today: multi-CSP (Oracle, CoreWeave, Microsoft, GCP, AWS, plus smaller neoclouds), multi-chip (Nvidia Blackwell as primary, with AMD, Cerebras for low-latency inference, and a proprietary chip in development with Broadcom), and multiple products at multiple price points. The strategic logic is opex over capex: CSPs absorb capital expenditure while OpenAI pays as revenue comes in. She acknowledges the shift toward build-to-suit (the SoftBank/Texas data center) introduces more capex, but frames it as the next phase once OpenAI achieves investment-grade credit standing and can access lower-cost debt. > *"A year ago people talked about the commoditization of the LLMs. And frankly it's gone the opposite because as you start building an agentic layer — the harness is what brings the context, the memory."* ## [29:32] OpenAI's ad business and strategy Calacanis closes with the advertising question. Two of the three greatest consumer businesses ever built — Google's search ads and Meta's social graph targeting — are ad-funded. OpenAI has been testing ads in the free tier. Is advertising the mechanism that makes AI access truly universal? Friar's answer is structured around two principles: ads must never alter which result the model surfaces (the answer must always be the best one, not the sponsored one), and an ad-free paid tier will always exist. Within those constraints, she's bullish on OpenAI's ad potential: the combination of Google-style high-intent signals ("I want cool shoes for the stage") with Meta-style demographic context, plus memory ("it knows I'm me, it knows I'm a CFO, it knows I'm a mom with teenagers"), produces a targeting surface more potent than either predecessor. The current API-versus-consumer token economics are stark — API revenue per token dwarfs consumer revenue per token by an order of magnitude. But Friar is deliberately not optimizing for today's ratio. The strategy is to build an AI infrastructure layer that functions like a utility, capable of serving consumers, small businesses, enterprises, and governments at global scale — and advertising is what pays for the consumer floor. > *"If you know Google and Meta had a baby, it would be ChatGPT — because what you have in Google search is very high intent... We have more than that because we have memory."* ## Entities - **Sarah Friar** (Person): CFO of OpenAI; former CEO of Nextdoor; Stanford trustee; guest - **Jason Calacanis** (Person): All-In host; LAUNCH founder; angel investor - **David Friedberg** (Person): All-In host; CEO of The Production Board - **David Sacks** (Person): All-In host; Craft Ventures; White House AI & Crypto Czar - **Sam Altman** (Person): CEO of OpenAI; cutting ribbon at Seline, Michigan data center during recording - **Jony Ive** (Person): Designer collaborating with OpenAI on an unrevealed consumer hardware device - **Denise Dresser** (Person): OpenAI's Head of Revenue, in seat since December - **OpenAI** (Organization): AI research and deployment company; 900M weekly ChatGPT users, $122B raised March 2026 - **Anthropic** (Organization): AI competitor; filed S-1 confidentially during the episode recording - **ChatGPT** (Software): OpenAI's consumer AI interface; 900M weekly active users - **Codex** (Software): OpenAI's AI coding agent; crossed 5M users in a single weekend; fastest-growing internally in GTM teams - **Cerebras** (Organization): Chip partner; online for OpenAI inference; noted for low-latency coding workloads - **Broadcom** (Organization): Semiconductor partner for OpenAI's proprietary chip development - **Nvidia** (Organization): Primary chip partner; Blackwell current; Vera Rubin targeted for next major training run; Feynman series next - **Oracle** (Organization): CSP partner; co-developing the Seline, Michigan 1-gigawatt data center - **Agentic AI** (Concept): AI systems taking multi-step autonomous actions; driver of both revenue growth and inference compute demand - **AI infrastructure layer** (Concept): OpenAI's strategic framing — single foundational model serving many interfaces, positioned like a utility

Benchmark GP Bill Gurley joins Jason Calacanis, David Sacks, and Chamath Palihapitiya (David Friedberg out this week) for a 95-minute session covering six fronts of the AI debate: Gurley's new theory that Anthropic is not just pursuing regulatory capture but actively "midwifing a deity"; Pope Leo XIV's 235-page AI encyclical and its uncomfortable historical parallel to Leo XIII's 1891 warnings about the industrial revolution; the growing consensus that open-source AI faces a coordinated regulatory crackdown; and the week's sharpest narrative flip — Dario Amodei and Sam Altman both quietly walking back their AI jobs-apocalypse rhetoric while Goldman Sachs CEO David Solomon published a New York Times op-ed declaring the apocalypse overblown. ## [00:00] Bill Gurley joins the show! Bill Gurley, Benchmark general partner and author of *Running Down a Dream*, fills in for David Friedberg and joins live from Chamath's pool house where Jason has been staying. After banter about unauthorized Uber Eats orders on Chamath's house iPad, Jason introduces Gurley as a first-time guest who specifically requested to appear the moment the pod covered the Pope. Gurley plugs his new P3 Institute and a grant program he launched to fund people pivoting toward work they love. He teases a TED talk — rooted in the book's argument that high agency and lifetime learning are the only durable defenses against disruption — which sets the frame for everything that follows. > *"And I told the house manager like, listen, any packages that come in the next 72 hours, right to the pool house, if it says JCAL, right to the pool house."* ## [06:00] Making yourself valuable in the age of AI, first class of "AI Natives" Chamath opens with the question that has been driving the show for 18 months: if you're a young person right now, is AI doom much ado about nothing, or a real career threat? Gurley cites a Gallup poll showing 59% of workers are "quiet quitters" — ambivalent about their jobs and therefore low-agency. His core thesis: the best protection against AI displacement is becoming the most AI-enabled version of yourself in your field. He invokes Mark Cuban's framing — "there are two types of people: those who use AI to learn faster than ever before, and those who use AI to avoid learning altogether." Sacks walks through how the pod's producer Nick built a daily Claude briefing document that not only summarized news but predicted specific topics Sacks would care about based on his prior comments on the show. Sacks had dismissed it as likely AI slop; it was not. Gurley extends the point across every job category: in marketing, legal, accounting, and sales, being the most AI-capable person among your peers makes you "golden," and the early lead compounds. Jason adds that in his own team experiments, the skill separating strong performers from weak ones was systems thinking — could they break a complex problem into context the AI could execute, or did they hand it a task and wait? > *"I think the best way to protect yourself from AI is to be the most AI enabled version of yourself you can be."* ## [17:37] Reacting to Pope Leo's AI encyclical: Who guards the guardians? Pope Leo XIV released *Magnifica Humanitas*, a 235-page, 42,000-word encyclical warning business leaders to safeguard humanity from AI. His central argument: technology is never neutral — it takes on the characteristics of those who build, finance, and control it. Jason reads the core line and notes the Pope presumably does not think highly of Silicon Valley's current roster of builders. Sacks finds himself largely agreeing with the Pope's diagnosis: the biggest risk of AI is centralization of power and its Orwellian misuse by governments. Where he parts ways is on the remedy. Giving government the power to regulate AI development creates its own guardian problem — the American founders' answer to *Quis custodiet ipsos custodes?* was separation of powers, forcing guardians to check each other. Sacks's AI equivalent: a competitive market with five frontier labs is the best natural check; monopolization is the scenario to prevent. Gurley lands the sharpest historical counterpunch. Pope Leo XIII's 1891 encyclical *Rerum Novarum* warned that the industrial revolution would harm workers — and was wrong on every metric. From 1891 to today: the work week fell from 60+ hours to 34, real wages rose 8–10x, the median worker now earns more than a doctor did in 1891, global GDP per capita went from $1,500 to $20,000, child labor in the US dropped from 18% to zero, workplace deaths fell 40x, life expectancy rose 60%, and global poverty dropped from 75% to under 10%. > *"All those things happened because of technology, innovation, and capitalism, which is exactly what Leo the 13th was warning against. So he got it dead wrong. He got the whole thing precisely wrong."* ## [26:54] Anthropic's Digital God: Do they believe they are creating a superior species? Gurley delivers what becomes the most-quoted segment of the episode: his "Dr. Frankenstein theory" of Anthropic. He had previously held a simpler regulatory-capture theory — Anthropic stirs up AI fear to lock in regulation that entrenches incumbents. But after spending 30 days reading everything he could find about the company, he has a darker read. He describes meeting people inside Anthropic who he believes genuinely think they are not writing software but "midwifing a deity." The evidence trail: Anthropic chief philosopher Amanda Askell's podcasts, Chris Olah's 80-page Constitutional AI document, and Dario Amodei's own essay "Machines of Loving Grace," which envisions a post-AGI economy where AI systems allocate resources to humans based on an AI-determined reward function. Chamath calls it "a computational reward function for humans — it decides how much you're worth." Jason calls it "the ultimate delusions of grandeur." Gurley corrects him: he didn't say it, Dario did. Sacks steelmans Anthropic briefly — they probably see themselves as responsible builders who take the power of this technology seriously enough to guard it — then immediately notes this framing is textbook regulatory capture: brand yourself the safe player, characterize competitors as reckless, let regulation shut down the recklessness. Both Sacks and Chamath converge on the structural danger: a singular AI value system that decides how humans live is catastrophically fragile. The answer is decentralization and competing systems, not one algorithmic authority. > *"I don't think they think they're writing software. I think they're midwifing a deity here. And I don't know which one I'm more afraid of — the regulatory capture or this second theory I call the Dr. Frankenstein theory."* ## [38:32] AI sovereignty, the next era of privacy, open-source crackdown coming? Jason introduces "intelligence sovereignty" as the successor to data privacy. Data privacy was about who can see your photos and messages. Intelligence sovereignty is about who gets to interpret your world — whether the AI shaping your information feed is a centralized system with a particular political philosophy, or something you control. He flags the paradox: China's Communist Party is leading the open-weight model movement while the United States is centralizing. Chamath presents his portfolio company Abacus as evidence that Fortune 1000 buyers are responding to this anxiety: they want a control plane that can hot-swap between frontier models, plus on-prem options that remove dependence on any one provider's terms of service. He gives a concrete example — a Canadian hospital that supports its country's euthanasia laws could be shut off by an American frontier model whose constitution prohibits that content. Sacks connects the dots to a regulatory threat he has been watching build: the regulatory-capture playbook leads, in his read, to a ban on open-source or open-weight models. The justification will be safety — open models let users strip guardrails. Gurley reaches the same conclusion in his P3 Institute post. If a ban succeeds, the United States effectively exiles itself from the open ecosystem while the rest of the world — including China — runs on open models. > *"I think where it's all leading to is an effort to ban open source models or open weight models. There's a lot of breadcrumbs leading here."* ## [59:56] The Great AI Jobs Debate: Dario and Sam Altman flip their rhetoric, Goldman CEO says no AI job apocalypse The chapter opens with a news roundup of the week's narrative shift. Cloudflare's Matthew Prince, Zuckerberg at Meta, Jack Dorsey at Block, and Andy Jassy at Amazon all cited AI when announcing major layoffs. But Goldman Sachs CEO David Solomon published a New York Times op-ed with three counterpoints: AI will automate 25% of work hours, not 25% of jobs; bank tellers increased after ATMs; the US labor market creates and destroys 25–35 million jobs annually so gross churn dwarfs net losses. Simultaneously, Fortune reported that Dario Amodei and Sam Altman are both walking back prior doom-and-gloom rhetoric — with Chamath noting the timing cannot be separated from upcoming frontier-lab IPOs that need a jobs-creation narrative. Sacks is unambiguous: he has been making the non-consensus case against the jobs apocalypse for over a year and considers himself vindicated. Yale Budget Lab found no discernible labor-market disruption over three years of the AI wave. Software engineering — the single breakout AI use case — saw job postings rise 15% year-over-year and hit a three-year high. The 4.3% unemployment rate is near record lows. Most of the high-profile layoffs, he argues, are AI washing: CEOs who over-hired during COVID found AI to be a convenient narrative for long-overdue downsizing. The Jack Dorsey / Block 50% cut was immediately flagged by financial analysts as a company that had been overstaffed relative to peers for years — pure AI washing. Jason pushes back. He insists cab drivers, truck drivers, and package-sorters — roughly 20 million American workers — face real structural displacement over the next decade regardless of current aggregate statistics, and accuses the panel of elitism: "We are elite performers. These people are going to lose their jobs and they may not get a job very quickly." He draws a distinction between the short-to-medium term, where he expects acceleration, and the long run, where a Cambrian explosion of startups built by AI-enabled founders creates new categories. By the end, he shifts toward Sacks's territory — acknowledging the aggregate data is less alarming than his anecdotes suggested. Gurley threads the needle with the same historical argument from the Leo XIII discussion: innovation has always, on net, created more prosperity than it destroyed. His practical advice to people at risk: get ahead of your peers on the tools now; if your job is going away, plan your pivot toward trades (he plugs MicroWorks, which provides free scholarships for plumbers, welders, and electricians) or toward something you find genuinely fascinating. > *"I think the best way to protect yourself from AI is to be the most AI enabled version of yourself you can be. Know what it's capable of in your field. Get out there."* ## Entities - **Bill Gurley** (Person): General partner at Benchmark; author of *Running Down a Dream*; founder of P3 Institute; guest filling in for David Friedberg - **Jason Calacanis** (Person): All-In host; angel investor; founder of LAUNCH; argues for worker empathy and short-term displacement risk - **David Sacks** (Person): All-In host; Craft Ventures founder; most vocal critic of AI jobs-apocalypse narrative this episode - **Chamath Palihapitiya** (Person): All-In host; Social Capital CEO; coined "intelligence sovereignty"; co-founder of Abacus - **Dario Amodei** (Person): Anthropic CEO; subject of Gurley's "Dr. Frankenstein theory"; walked back jobs-doom rhetoric this week alongside Sam Altman - **Pope Leo XIV** (Person): Catholic Pope; released *Magnifica Humanitas*, a 235-page AI encyclical warning against technology concentration - **David Solomon** (Person): Goldman Sachs CEO; published New York Times op-ed arguing AI job apocalypse is overblown - **Anthropic** (Organization): Frontier AI lab; subject of Gurley's regulatory-capture and "Dr. Frankenstein" theories; maker of Claude - **P3 Institute** (Organization): Bill Gurley's new policy and philanthropy institute; published post defending open-source AI - **Goldman Sachs** (Organization): Investment bank; CEO's NYT op-ed became the week's anchor data point against the jobs-apocalypse narrative - **Abacus** (Software): Chamath's Social Capital portfolio company; builds on-prem AI hardware stacks for Fortune 1000 enterprises seeking model independence - **Intelligence sovereignty** (Concept): Jason's term for the next frontier of privacy — not who sees your data, but which AI system is allowed to shape your interpretation of the world - **Dr. Frankenstein theory** (Concept): Gurley's characterization of Anthropic's worldview: senior staff believe they are midwifing a deity or superior species rather than writing software, as described in Dario Amodei's "Machines of Loving Grace" essay - **Regulatory capture** (Concept): The strategy of branding oneself the "safe" AI company, amplifying public fear, and lobbying for regulation that locks in incumbents and targets open-source competitors

Sacks is out, Gavin Baker (Atreides Management) sits in. The panel walks through Andrej Karpathy's surprise move to Anthropic, debates why the public mood on AI has flipped, tears apart SpaceX's $2T S-1, and asks why Nvidia's blowout earnings still saw the stock sold. Friedberg and Chamath also flag warning signals from inflation, oil, and bond yields, and close on what — if anything — came out of the US-China summit. ## [00:00] Gavin Baker joins the show! Jason opens episode 274 noting Sacks is out and welcomes Gavin Baker from Atreides Management for the week. They tee up the agenda: SpaceX and OpenAI IPOs, Karpathy to Anthropic, and Nvidia's earnings. > *"Sachs is out today, but we're very lucky to have Gavin Baker from Atreides Management joining us. The spicy takes must flow."* ## [00:30] Andrej Karpathy joins Anthropic; hypergrowth and profitability The Karpathy hire is read as a major strategic win for Anthropic — Chamath frames it as continuity of the Richard Sutton "bitter lesson" school of scaling that Karpathy executed at Tesla FSD and OpenAI. Gavin layers in financial context: Anthropic was EBIT-positive in the last quarter per the WSJ, which combined with hypergrowth makes the recent funding rounds look very different from a capital-burn narrative. Friedberg pushes back on the framing that models will soon "feed themselves" into context windows to self-improve, but flags that papers (one from MIT) suggest large efficiency gains are on the horizon. Chamath uses the moment to argue the podcast itself has to start telling the upside story of AI — the doctors, the scientists, the unlock — because the dominant public narrative has gone negative. > *"He was probably the first person that really commercialized the Richard Sutton bitter lesson essay when he was leading FSD at Tesla."* ## [12:42] Why Americans have turned on AI, anti-human perception Gavin shares a personal story: his daughter has a rare disease, and a Stanford scientist he funded is months away from what he believes is a complete cure, made tractable by AI-accelerated biology. He uses it to argue for an optimistic posture — a future where work is optional and disease is solvable — and warns that the people pushing for AI regulation are also shaping how the public feels about the technology. Friedberg goes deeper into the cultural mechanics: AI is being framed as anti-human in a way that mirrors anti-nuclear and anti-industrial backlashes of the 20th century. He argues the United States can't unilaterally slow down because China and others won't — and tries to separate genuine safety concerns from elite class anxiety. Chamath then makes a pointed observation that none of the survey data on AI job loss actually asks the truck drivers, package sorters, and ICU nurses themselves how they feel about the tools. > *"We're listening too much to the inventors of AI. They're geniuses. They're smart. We need to be listening to the frontline factory workers who are using AI saying, 'Wow, I was able to add a third shift.'"* ## [27:22] Trump pulls AI EO, US-China AI relationship, dystopian AI layoffs A Trump AI executive order was scrubbed at the last minute — the panel walks through what was reportedly in it (review of frontier-model training runs) and whether any pre-release regulatory framework is workable. Jason argues a state-by-state patchwork is the more likely outcome regardless of what Washington does. The conversation pivots to Meta's latest round of layoffs and the way they were communicated. Gavin and Jason agree the messaging — leaning on "AI productivity gains" as the public reason — landed badly even with people who accept the underlying logic, and Jason argues it became a case study in how *not* to message AI-driven workforce changes. > *"Because the reality is that if this is the way that you're going to message something as critical as this, I think you did a horrible job."* ## [45:19] SpaceX S-1 tear down! Breaking down the three major businesses and the case for a $2T valuation SpaceX filed its S-1 on Wednesday. Jason breaks the company into three businesses: launch (which could be hundreds of millions of paying subscribers via Starlink), Elon Web Services / xAI / Colossus compute, and rockets. The AI-cloud line item alone is around $15B and growing roughly 2x year over year, anchored by an Anthropic deal Gavin calls "extraordinary." Gavin then makes the case that Colossus matters because raw gigawatt-class data centers are now the binding constraint, and SpaceX-adjacent build velocity is the moat. He uses Cursor's Composer 2.5 release — Pareto-dominant on three or four weeks of RL training — as evidence that whoever owns the compute owns the next model generation, and walks through why rapid reusability on Starship compresses the unit economics of getting payload to orbit faster than any competitor can model. > *"If you look at who's actually capable of delivering a gigawatt data center, these guys are the closest, like an actual gigawatt."* ## [71:22] Nvidia smashes earnings but stock falls, why people are shorting chips Nvidia blew out earnings again — 20% sequential growth would be a high-growth print for any other company, the dividend was raised 25x, and the CFO committed to returning 50% of free cash flow. Yet the stock sold off, and Leopold Aschenbrenner's reported pivot away from chip exposure is being read as a smart-money signal. Gavin takes the bear case apart: at current PE Nvidia is cheap relative to growth, and the segment breakdown obscures how much the "AI clouds" line is dragging the multiple. He flags that the true useful life of a GPU is closer to two years than five, which means the reported profits of every hyperscaler running these chips are overstated — a real concern, not a stock-killer. He also notes Nvidia's CPU business is on track to do $20B this year, making it overnight one of the largest CPU manufacturers in the world. > *"The true lifespan of a GPU is more like two years and therefore the profits of all these businesses are overstated."* ## [82:25] Market update: Flashing red signals, oil, inflation, yields up The macro snapshot: May inflation expected at 4.2%+, Fed rate-hike odds back on the table, UK yields at the highest since the great financial crisis, oil and gold both moving. Chamath warns that when the currency-debasement mechanism finally breaks, the downside is non-linear. Gavin counters with relative optimism on the US: America is self-sufficient in energy, the AI build-out is structurally good for re-industrialization, and even in an ugly global scenario the US is the least-bad place to be invested. He flags AI fundamentals also have a seasonality that investors are starting to model — the same way e-commerce and subscription businesses do. > *"While it's terrible for everyone, it is relatively the best for America because we are self-sufficient in energy."* ## [92:45] China trip flops, or was progress made behind the scenes? A 48-hour US tech-CEO-plus-president trip to Beijing produced thin public deliverables: some soybeans, some H100/A200 sales to Chinese players. The panel asks whether that's the real story or just the visible surface, and whether the immediate China-Russia bonding moment afterward says more about the trajectory than any handshake photo. Gavin argues the more important read is structural: keeping America ahead in AI requires keeping the trans-Pacific relationship just stable enough to avoid a full decoupling shock, and that's a defensible strategic logic even if the optics are unsatisfying. He also paints a what-if scenario around the Strait of Hormuz to make the point that energy independence is what gives the US the option to act asymmetrically. Jason closes with thanks to Gavin and an invite back to the Summit. > *"There's sound arguments that this is stabilizing for the world and is the best highest probability path for keeping America ahead in AI."* ## Entities - **Jason Calacanis** (Person): Host, LAUNCH founder, MC of this episode. - **Chamath Palihapitiya** (Person): Host, Social Capital CEO; pushed the "listen to frontline AI users" framing. - **David Friedberg** (Person): Host, The Production Board CEO; led the cultural / historical analysis of the AI backlash. - **Gavin Baker** (Person): Guest host, Atreides Management founder/CIO; carried the investing thread across SpaceX, Nvidia, and macro. - **Andrej Karpathy** (Person): Joining Anthropic's new pre-training team; OpenAI co-founder, ex-Tesla FSD lead. - **Anthropic** (Organization): Hired Karpathy; EBIT-positive last quarter per WSJ; $15B AI-cloud deal with SpaceX-adjacent compute. - **SpaceX** (Organization): Filed S-1; three businesses (launch/Starlink, Elon Web Services compute, rockets); $2T valuation case. - **Nvidia** (Organization): Earnings blowout but stock sold off; $20B CPU run-rate; $5.3T market cap. - **Cursor** (Software): Composer 2.5 model release used as proof of fast RL-driven catch-up dynamics. - **Richard Sutton's bitter lesson** (Concept): Scaling beats clever architectures — framing for why Karpathy's move matters. - **GPU useful life** (Concept): Closer to ~2 years than ~5, so hyperscaler reported profits are overstated. - **Strait of Hormuz scenario** (Concept): Energy-independence-as-strategic-option argument for the US in the China game.

SalesforceのCEOマーク・ベニオフが、Jason Calacanis、David Friedberg、Chamath Palihapitiya（David Sacksは欠席）とともに、二つのリアルタイムニュースを軸に幅広いテーマを語り合う。トランプ・習会談（2017年以来初）と、企業向けソフトウェアの評価額を急速に侵食するAIだ。サウジ国賓晩餐会、ウィンザー城、今回の代表団にも同席したベニオフは、米中経済外交の最前線を伝える。その後、話題はSalesforce自身の株価再評価の本質へ移り、データインフラとエージェントプラットフォームがAI破壊の恩恵側に立つと主張する。後半はOpenAI対Apple、Thinking Machinesのリアルタイムマルチモーダルデモ、Friedbergが提示する衝撃的なエルニーニョデータ、そしてAnthropicによる多層SPVスキームへの取り締まりを扱う。 ## [00:00] Salesforce CEO マーク・ベニオフ、番組に登場！ 今週はSacksが不在で、ベニオフが席を埋める。Jasonがさっそくベニオフの政治的立ち位置を問いただす——かつて民主党の資金提供者だった男が、今やサウジ国賓晩餐会に出席し、現政権にも受け入れられているのはなぜか。ベニオフは党派的な問いをきっぱり退けた。 > *「私は民主党員でも共和党員でもない。アメリカ人だ。」* Chamathは、ベニオフがウィンザー城、チャールズ皇太子の訪米、サウジ国賓晩餐会への招待を立て続けに手にした珍しい存在だと指摘する——どの政権とも摩擦なく動けるテックCEOだ。このくだりは、まさに進行中の会談について発言する資格がベニオフにあることを際立たせる。 ## [01:14] トランプ・習会談、米企業の中国ビジネス、米国民と中間選挙への影響 トランプと習近平の七回目の直接会談——イラン戦争で二ヶ月遅延——は北京で幕を開け、習が台湾の扱い次第では両国関係が「極めて危険な状況に陥る可能性がある」と警告した。Polymarketでは2026年の侵攻確率が2300万ドルの出来高をもとに6%と示された。貿易面では、習が大豆・米国LNG・ボーイング機200機の購入を約束し、商業面での「門戸拡大」を訴えた。米国側代表団は企業役員会の顔ぶれだ——Jensen HuangはチップをPR、Kelly Ortbergは航空機を売り込み、CargillのBrian Sykesは大豆を売り、VisaとMastercardは決済市場へのアクセスを求めた。 Friedbergは「修辞ディレス・トラップ」の視点から会談を読み解いた——台頭する大国と衰退する大国が衝突するのは歴史の定型だが、AIやバイオテクノロジーがもたらす資源膨張という稀有な瞬間に、そのパターンから抜け出す可能性があると主張した。 > *「AIや自動化、バイオテクノロジーによる技術の大転換が目の前に広がり、真の豊かさが見えてきたこの瞬間こそ、世界がより多極的な在り方を選べる絶好のタイミングだと思う。」* ベニオフは、Salesforceが中国本土にオフィスも従業員も持たないことを明かした——中国の収益はすべてデータ所在地法を満たすためのAlibaba独占パートナーシップ経由だという——そして代表団全体に実際の受注が生まれると期待を示した。Chamathは、中国のトップダウン型儒教的ヒエラルキーがCEOレベルの外交を官僚チャンネルより効果的にすると論じ、インフレに苦しむ米国民にとってもこの取引成立が不可欠だと訴えた。 ## [18:46] 台湾、半導体、AIモデル、貿易による平和 ベニオフは、台湾が習にとって最優先事項だという前提に異を唱えた。領土的野心より経済的繁栄と中産階級の成長の方が習にとって重要だというのが彼の立場だ。「もし中国が台湾を封鎖したら米国は守るべきか」という直球の問いに対しては、二者択一を拒みこう述べた。「中国と台湾は和解すると思う。」Chamathは構造論で応じた。米国は国内チップの同等性まであと1〜2ナノメートルという地点にいる。そこに達した時点で、台湾の戦略的価値は存亡に関わるものから経済的なものへと変わると見た。 > *「私たちは、台湾が戦略的に担っている役割を自力で果たせるまで、あと1〜2ナノメートルのところにいる。今は経済的な問題であって、それが解消されれば、台湾に対する態度はまったく変わると思う。」* Chamathの処方箋は明快だ——それでもチップを売れ。HuaweiがKYC管理なしに半導体レースを制するよりも、NvidiaがKYC付きで中国にモデル用途のチップを販売する方がましだ。ベニオフも、チップ規制にもかかわらず中国のAIモデルが米国モデルとほぼ同等の性能に達していると同意し、禁輸の根拠を自ら崩した。Friedbergは、中国が国内ファブと製造装置を整備するにつれ、政治的な結果に関係なく台湾の代替不可能性は自然と低下していくと付け加えた。 ## [31:41] AIがソフトウェアに与える衝撃：生き残るSaaSと死ぬSaaS Jasonは株価再評価を率直に突きつけた。Salesforceが37%、ServiceNowが42%、Workdayが45%下落——AIがマネージドSaaSを不要にするという見立てのもと、合計で約1800億ドルの時価総額が消えた。ベニオフは真っ向から反論した。 > *「正直、SaaS崩壊は今回が初めてじゃない。でもこれが今のSaaS崩壊だ。」* 彼の主張はこうだ——市場の再評価は誤った前提に基づいている。Salesforceが賭けているのはAgentforce、つまり実際の企業データに根ざしたAIエージェントであり、幻覚を起こしやすい汎用モデルではない。80〜90億ドルでのInformatica買収が提供するデータ統合レイヤーこそ、エージェントを信頼できるものにする。「AIは確率論的だから、真実に、一元的な真実の情報源に縛り付けないと、うまく機能しない。」ベニオフはさらに、Salesforceが今年内部コーディングエージェントのためにAnthropicに約3億ドルを使う予定だと述べ、実装サイクルが劇的に短縮されるとした。 Chamathは市場を二つに割った。ローエンドはもう終わりだ——深い顧客関係を持たない汎用のポイントソリューションは死んだ。だが、Salesforceが戦うハイエンドは、公開市場が「AIに浮かれる」のをやめてCapex3兆ドルが何を生んだかを問い始めたとき、ROIの精算から恩恵を受ける側に立つ。生き残るのはCスイートとの関係、ネガティブチャーン、そしてAI機能を測定可能な成果としてパッケージできる企業だ。 ## [47:26] OpenAI、ChatGPT統合失敗をめぐりAppleを提訴か Bloombergによると、OpenAIがAppleを契約違反で提訴する可能性がある。2024年のChatGPT・Siri連携は実態として機能しなかった——Appleはユーザーが明示的に「ChatGPT」と言った場合だけクエリを転送し、統合を積極的に宣伝せず、OpenAIは期待していたサブスクリプション収益を得られなかった。Appleの言い分はOpenAIのデータ管理への懸念だ。 ベニオフはこの話をAIラボ間の戦略的な分岐として捉え直した。Grokはコンパニオンアプリやいわゆるセックスボットを作り、OpenAIはSoraと広告ネットワークを推し進め、GeminiはNanoを出荷した。そしてAnthropicはそのすべてを無視してコーディングエージェントに集中した——それが正解だったと彼は言う。まだ発表されていないSlack内蔵のコーディング機能についても示唆した。 > *「Anthropicはそんなセックスボットは知らない、Nano Bananaも知らない、コーディングエージェントをやる、と言った。結果的にAnthropicが正しかった。突然、ロケットが飛び上がった。」* Chamathはより深い問いを立てた——AI対話レイヤーがデバイスから完全に離れたとき、Appleに何が起きるのか。彼は、予想外のハードウェアメーカーからMacBook Proを不要にするような薄型・常時起動のアンビエントデバイスが「iPhoneモーメント」をもたらすと予測した。FriedbergはAppleの現在の戦略は先見性ではなく空白の穴埋めだと指摘し、G Suiteが静かにAppleの生産性スタックから企業シェアを奪いつつあると述べた。 ## [56:54] Thinking Machinesがリアルタイムモデルを公開、コンシューマーAIの未来、マルチセンサーモデル Mira MuratiのThinking Machinesが、デスクトップを監視し、周囲の音声を聞き取り、ウェブカメラの映像を同時処理するリアルタイムマルチモーダルモデルを公開した。200ミリ秒間隔で二本の並列パイプラインが走る——一方は深い遡及的推論、もう一方はライブ応答用だ。Appleはこれと並行して、AirPods内蔵カメラの特許を出願している。 > *「マルチセンサーモデルはAIの次の大きな波だ。そこに達してもまだAGIではないけれど。」* ベニオフは、言語で訓練されたLLMには根本的な限界があると論じた。人間の認知は目・耳・固有感覚を生体ハードウェア上で並行処理している。マルチセンサーによる根拠付けこそが欠けているレイヤーだ。トークン経済のインパクトは大きい——一日8時間のリアルタイムアンビエント監視は、現在の企業用途の1000倍の消費量になる。ベニオフは「大きいモデル＝良いモデル」という軍拡競争に疑問を呈し、アプリやデバイスに組み込まれた分散型インテリジェンスがモデルの生のスケールより重要になると予測した上で、アンビエントセンシングと企業コンテキストを統合する「熱い新興企業」の余地があると示唆した。 ## [62:24] サイエンスコーナー：2026年に記録的規模のエルニーニョが起きたら Friedbergが海面水温の異常データを提示した。1877年以来最大の偏差に向かっており、基準値から約4℃上昇している。蓄積された熱エネルギーは1100万テラワット時——人類の年間エネルギー消費25,000テラワット時と比べると桁が違う。 > *「この海には500年分の人類エネルギーが蓄えられている。今後数ヶ月でそのエネルギーが大気中に放出される——99%の確信を持って言えるが、来年は記録上最も暑い年になり、しかもダントツになる。」* 連鎖反応はこうだ。貿易風の変化がカリフォルニアとメキシコ湾岸に大気の川をもたらし、熱ドームがフェニックスとカナダ内陸部に広がる。インドのモンスーンが高確率で失敗し、1億5000万人の農家と食料を依存する15億人が危機にさらされる。ブラジルのインドネシア・フィリピン向け農作物輸出が崩壊し、小麦価格が世界的に急騰する。フェニックスはすでに5月に摂氏41℃（106°F）を記録した。コモディティ市場はエルニーニョへの露出を積極的に取引している。Friedbergが示す部分的な好材料は、作物の遺伝子改良で干ばつ耐性が向上したことと、シベリアの農地が拡大していることだ——ただしそれらが2026年の収穫期を救うには間に合わない。 ## [71:40] Anthropic、「ダーク SPV」に宣戦布告 Anthropicは、非上場AI企業の未公開株を個人投資家に販売する多層SPVプラットフォームを公式に問題視し、許可を受けていない仕組みを通じて売却された株式を無効にすると表明した。Chamathは全面的に支持した。上場前のすべての企業が同じ姿勢を取り、公開市場への移行を押し進めることで、こうした仕組みを消滅させるべきだと主張した。 > *「SpaceXが上場し、Anthropicが上場し、OpenAIが上場すれば、これらSPVの販売業者を巻き込んだ訴訟の嵐が始まるだろう。こんな仕組みは許されるべきではない。」* Chamathは、主要なAI企業が上場し、個人SPV投資家が計算が合わないと気付いた時点で大量の法的紛争が起きると予測した。最後にベニオフが1-1-1慈善モデル——創業時に株式1%・利益1%・従業員時間1%を拠出し、現在は5万の非営利団体がプラットフォームを無料で使っている——とSusan Wojcickiへの追悼を語り、章は締めくくられる。 ## 登場人物・概念 - **Marc Benioff**（人物）: SalesforceのCEO兼会長。本エピソードのゲスト。1-1-1慈善モデルおよびAgentforce AIエージェントプラットフォームの生みの親 - **David Friedberg**（人物）: ホスト。The Production BoardのCEO。エルニーニョのサイエンスコーナーを担当 - **Chamath Palihapitiya**（人物）: ホスト。Social CapitalのCEO。SaaS高付加価値領域の生き残り論とNvidiaチップ普及論を展開 - **Salesforce / Agentforce**（ソフトウェア）: 企業向けCRMおよびエージェントプラットフォーム。データ根拠型AIエージェントがSaaS終焉論の裏返しになるというベニオフの賭け - **Anthropic**（組織）: AIセーフティ企業。Benioffが好むコーディングエージェントの提供元（Salesforceで年間約3億ドルの支出を予定）。無許可SPV構造への取り締まりも実施中 - **OpenAI**（組織）: ChatGPT・Siri統合の失敗をめぐりAppleを提訴する可能性があると報道。Anthropicの成功を受けてコーディングエージェントに軸足を移しつつある - **Thinking Machines / Mira Murati**（組織）: デスクトップ・音声・ウェブカメラを200ミリ秒間隔で同時処理するリアルタイムアンビエントマルチモーダルモデルを公開 - **Thucydides Trap（修辞ディレス・トラップ）**（概念）: 台頭する大国と衰退する大国の衝突サイクルを指す政治学の枠組み。Friedbergが米中会談における協調的な豊かさの好機を語る文脈で引用 - **ダーク SPV**（概念）: 非上場AI企業の株式を個人投資家に販売する多層の特別目的ビークル。高い手数料と法的根拠の曖昧さが問題視されている

Charles KochとChase Kochが、David Friedbergと共に振り返るのは、Koch Inc.が9,000倍の成長を遂げた歩みだ——1961年のオクラホマの300人規模の石油会社から、エネルギー・化学品・森林製品・消費財・ベンチャーキャピタルにまたがる13万人のプライベートコングロマリットへ、一度も上場せずに成長した。対話の中心は原則に基づく経営（PBM）：Koch のすべての採用判断・買収・文化変革を動かす41の原則フレームワーク。Charles と Chase は「Koch」という名前につきまとう狭い政治的イメージについても率直に語り、党派的なリバタリアン政治から教育改革と人間の繁栄に焦点を当てるStand Togetherコアリションへの転換を説明する。エピソードはAIと資本主義の議論で締めくくられ、両者はパーミッションレス・イノベーションとボトムアップの権限委譲こそが、これからの経済的圧力を乗り越える唯一の信頼できる道だと主張する。 ## [00:00] David FriedbergがCharles & Chase Kochを迎える David Friedbergは、Forbes主催のイベントでの対話を開幕し、Chase Kochとは農業産業で2013年から知り合いでビジネスパートナーでもあると述べる。Koch Inc.をアメリカのビジネス史における「語られてこなかった物語」と位置づけ、おそらく世界で最も収益性の高いファミリー経営の非公開企業でありながら、上場企業と比べてほとんど知られていないと語る。 また、今夜の目的を明示する——Koch Inc.の会長と次世代の社長の両者との希少な長時間のライブ対談を、All-Inの視聴者に届けること。 > *「Koch Industriesはまさに語られてこなかった物語だと、ずっと感じていました。おそらく世界で最も収益性の高いファミリー経営の非公開企業だと思います。」* > — David Friedberg ## [01:04] Koch Inc.の概要：規模・事業・歴史 Friedbergが統計値を提示する。もしKoch が上場企業であれば、収益規模でFortune 500のトップ25に入る。1940年にFred Kochがウィチタで創業し、現在は60カ国でエネルギー・農業・化学品・建材・消費財・クラウドコンピューティング・マイノリティ投資ポートフォリオにまたがる12万人超の従業員を擁する。Koch は利益の90%を事業に再投資している——四半期収益を最大化しようとする上場企業とは異なる構造的な選択だ。 Charles は、今夜の対話が本当に何について語るかを示唆する。収益の節目の話ではなく、持続的な複利成長を可能にした原則と失敗の話だ。 > *「破壊的イノベーション、利益の90%を新事業と成長に再投資する姿勢、能力主義的な価値観を含む、非常にユニークなオペレーティングモデルです。」* > — David Friedberg ## [02:21] ビジネスの構築：初期の歩みとCharles Koch入社（1961年） Charles Kochは、MITとArthur D. Littleでの経営コンサルタント経験を経て25歳で家業に入社した1961年を振り返る。父Fred Kochの言葉は明快だった。「お前が戻って会社を経営してくれなければ、売却しなければならない。体調が悪く、会社も調子が悪い。もう長くないんだ。」当時は約300人の従業員と、フラクショネーティングトレイ製造とオクラホマの原油集積という二つの主力事業があり、運営は機能不全の状態だった。 初期の経験から一つのKoch 原則が結晶化した——業界にとらわれず、ケイパビリティを軸に成長するということ。フラクショネーティングトレイ事業が失敗した一因は、社長がトップダウン型のコントロール主義で、エンジニアと顧客の双方を疎外したことにあった。Charles はやがて「どの業界にいるか?」という問いではなく「誰よりもうまくできることは何か、バリューチェーンのどこでそれが最も価値を生むか?」と問うようになった。この発想の転換を何十年にもわたって繰り返すことが、Koch がのちに無関係に見える多様な産業に進出していく理由を説明する。 > *「息子よ、戻って会社を経営してくれなければ、売却しなければならない。私は体調が悪く、会社も調子が悪い。もう長くないんだ。」* > — Charles Koch、父Fred Kochの言葉を引用 ## [11:31] 失敗・創造的破壊・ミスからの学び Charles は挑発的な言葉で口を開く。「すべてにおいて失敗していなければ、新しいことは何もしていない証拠だ。」石油コークスから活性炭を製造しようとした失敗など、初期の損失の数々を振り返る。失敗のほぼすべてに共通するのは、Koch の経営原則のどれかを違反していたことだ。 Chase はケイパビリティ・ポートフォリオの視点を加える。原油集積から天然ガス・化学品・肥料・そして最終的には森林製品へと事業を広げたのは、無秩序な多角化ではなく、同じ根本的なケイパビリティを新しい用途に向けたものだ。また自身が設立したKoch Disruptive Technologies（KDT）を「構造的に継続的な収益を上げにくかった」と正直に評価する。撤退・ピボットの判断は一つのテストに尽きる——顧客に対して優れた価値を創出し、そして報われる能力を失っていないか? > *「十分に大損した時——それが「もう十分」のサインだ。顧客に対して優れた価値を創出できないとわかった時だ。」* > — Charles Koch ## [19:22] 文化と原則に基づく経営（PBM） このエピソードの知的な核心部分だ。Charles はPBMの源流を、Koch 最大の失敗——すべてに共通の根本原因、つまり悪い価値観を持つ人間をリーダーに登用したこと——にまで遡って説明する。二つの危機的事例が際立つ——1973年の中東戦争時に会社を倒産寸前に追い込んだ無謀なトレーディング、そして「破壊的な動機」を持つリーダーが失敗を隠し成功を捏造した後年の事件。解決策は、価値観を第一に、才能を第二に採用し、「貢献意欲」——他者の成功を助けることで自分も成功したいという動機——が権力志向を駆逐する文化を構築することだった。 Chase はさらに核心をつく問いを立てる——会社の全員が言われなくても何をすべきかわかるとしたら? それがPBMの目指す状態だ。変革戦略はトップダウンの指令を避ける。原則を試す意欲が最も高いグループを見つけ、成果を実証し、変革への需要が組織の残りの部分を引っ張るようにする。集合的な知識がトップの少数の優秀な人々の判断を置き換える。 > *「大小に関わらず、全員が言われなくても何をすべきかわかるような企業と文化を持てたらどうなるか?」* > — Chase Koch ## [33:53] Georgia-Pacific買収と文化変革 2005年のGeorgia-Pacific買収は当時の最大の賭けだった——Chase が言うように「当時の会社規模からすれば巨大な賭け」だった。Charles は論理をこう辿る。Koch は化学的プロセス産業と木材・パルプの「相互利益の好循環」を発見した。その連鎖はFred KochのMIT学位論文にまで遡る。最初はコモディティ部門のみの買収を提案したが、係争中の訴訟でそれが実現できないとわかり、会社全体を買収することになった。 その後は、トップダウン官僚制が支配していたアトランタ本社51階建てビルで、数年にわたる文化変革が始まった。Koch はリーダーを入れ替え、非効率を発見・解消した従業員を表彰し、それを見つけた組合員と節約分を共有した。Chase は自らが現場で過ごした年月——飼育場のシングルワイドトレーラーでの生活、ガスリキッドプラントでの仕事——が後年の信頼性ある経営の礎になったと語る。文化変革は買収側が想定するよりはるかに長くかかり、ほぼ必ず古いパラダイムを持つリーダー層の交代が必要だ。 > *「文化を変えるのは思っているよりずっと時間がかかる。そしてほぼすべての場合、ボトムアップ権限委譲のパラダイムを持つリーダーへの交代が必要です。」* > — Chase Koch ## [56:17] 教育改革と社会変革 Stand Together——Charles が60年にわたってさまざまな名称で構築してきた非営利ネットワーク——は今や米国最大の慈善団体の一つだ。Chase はoriginationとpartnershipsを担当し、その使命を再定義する。政治的擁護ではなく、同じKoch 原則を社会課題に適用すること、まず教育から。COVID-19は世論を大きく変えた。2020年以前は約20%の家庭が従来の学校教育の代替に開放的だったが、子供たちがZoomの授業よりYouTubeでより多くを学んでいる現実を目の当たりにした後、その数字は急増した。Stand Togetherはその後5,000校超のマイクロスクールの設立を支援してきた。 Joe Limontのアルファスクールのようなパートナープログラムは、ゲーミフィケーションとプロジェクト型学習を使って、失敗していた生徒を3ヶ月でトップクラスに変える。Chase は自身にも比較優位の原則を適用した——Koch Fertilizer社長として比較優位がないと気づき自ら辞任した——そして同じ視点でKoch 13万人の従業員の役割再設計に取り組んでいる。 > *「COVID以前は約20%の家庭が新しい教育モデルを受け入れる意思がありました。COVID中に皆が教育システムの問題を目の当たりにして、子供たちが教室のZoomよりYouTubeでずっと多くを学んでいることに気づいた。」* > — Chase Koch ## [72:37] AI・経済的課題と資本主義の未来 FriedbergはCharles に、Koch の政治的物語——リバタリアン党への長年の関与とStand Togetherの広範なコアリションへの転換——について問う。Charles は率直だ。最初の50年間、あらゆる原則で自分と同意する人たちとだけ仕事をしていたため、影響範囲が限られていた。Viktor Franklの洞察——「今日の問題は、ますます多くの人が生きる手段を持ちながら、生きる意味を持たないことだ」——が、純粋に政治的な解決策ではなく社会的崩壊の動機的根源へと思考を向けなおした。教訓：自由の戦略は全体主義から借りることはできない。コアリション内の純粋性テストはそれを破壊する。 AIについてChase の立場は明確だ。パーミッションレス・イノベーション、オープンシステム、AIツールで人々を力づけること——禁止するのではなく。Koch はPBMをAIネイティブなフレームワークとして運営しており、Chase は読者が原則と直接対話できるAIコンパニオンを本と合わせて作った——Charles が想像していた以上の展開だ。エピソードはCharles の遺産目標で締めくくられる——米国が独立宣言の約束をより完全に実現すること。 > *「今日の問題は、ますます多くの人が生きる手段を持ちながら、生きる意味を持たないことだ。」* > — Charles Koch、Viktor Franklを引用 ## 登場人物 - **David Friedberg** — ホスト、The Production Board共同創業者、農業産業を通じて2013年よりChase Kochとのビジネス上の知人 - **Charles Koch** — 1967年よりKoch Inc.会長兼CEO、MITで工学を学んだエンジニア、原則に基づく経営の本の共著者、Koch の9,000倍の価値成長を牽引 - **Chase Koch** — Koch Inc.社長、Koch Disruptive Technologies設立者、Charles との PBM本共著者、Stand Togetherのoriginationとpartnershipsを担当 - **Koch Inc.** — ウィチタKS本社のプライベートファミリーコングロマリット、1940年Fred Koch設立、エネルギー・化学品・森林製品・消費財・ソフトウェア・ベンチャーキャピタルにまたがる13万人超の従業員 - **原則に基づく経営（PBM）** — Koch の41の原則からなる経営フレームワーク、貢献意欲・価値観優先採用・ボトムアップ権限委譲・各事業を実験室として扱うことを重視 - **Georgia-Pacific** — 2005年にKoch が買収した森林・消費財企業、Koch 最大の買収案件、PBMによる文化変革の主要事例 - **Koch Disruptive Technologies（KDT）** — Chase Kochが創設したベンチャー部門、破壊的テクノロジー企業へのマイノリティ投資、構造的に継続的な収益を上げにくかったと評価 - **Stand Together** — 2003年よりCharles Koch の慈善ネットワーク、教育改革・貧困削減・党派を超えた社会変革に注力、COVID後に5,000校超のマイクロスクールを支援

In one of their most consequential episodes, the All-In besties dissect SpaceX's surprise compute lease to Anthropic — the deal that may cement Anthropic as AI's dominant platform — and debate whether David Sacks's "Rockefeller" framing is prophecy or paranoia. The group then wrestles with a White House trial balloon about an "FDA for AI," ultimately concluding it was mostly media spin, before closing with a bullish-but-cautious read on the AI-driven market boom. Brad Gerstner fills in for David Friedberg, bringing investor perspective from both public and private markets across the episode's 82 minutes. ## [00:00] Bestie intros! Thoughts on the LA mayor election Jason Calacanis opens with the full crew: Chamath Palihapitiya, David Sacks, and fifth bestie Brad Gerstner joining in for David Friedberg, who is out sick. The warm-up quickly turns to the LA mayoral race, where Spencer Pratt is mounting a surprisingly effective challenge to incumbent Karen Bass. The group praises Pratt's viral debate performance — evisceration of the city council candidate over homeless policy — and Chamath notes the power of a sharp social-media team in modern politics. Brad flags a California ballot initiative that would constitutionally protect retirement savings and ban a wealth tax, reading it as a potential seismic signal. Jason observes that New York City hedge-fund titan Ken Griffin publicly announced he is pulling investment from New York after NYC councilman Zohran Mamdani targeted his home in a campaign video, underlining the tension between aggressive progressive politics and capital flight. > *"If California effectively passes a constitutional amendment protecting retirement savings and personal assets and banning the wealth tax and [Spencer Pratt] gets elected, the message that would send to the country — that's a very non-consensus view that I'm becoming increasingly optimistic about."* — Brad Gerstner ## [04:38] SpaceX-Anthropic deal, Elon Web Services, SpaceX IPO valuation, Anthropic's insane growth trajectory Jason leads with the blockbuster news: SpaceX has leased all of Colossus 1 — its H100-based Memphis data center — to Anthropic, adding over 220,000 Nvidia GPUs and 300 megawatts to Anthropic's supply-constrained capacity. The deal immediately doubled Claude Code's rate limits and removed peak-usage caps for paid users. Chamath frames Anthropic's explosive growth as purely supply-constrained: if unlimited power existed, revenues would be "even more parabolic." He sees the deal as Elon strategically de-risking SpaceX's valuation story — blunting bear cases around delayed orbital data centers while generating near-term revenue to subsidize Grok training. Brad estimates the arrangement adds $4–5 billion in incremental 2026 revenue for SpaceX, calling EWS (Elon Web Services) a genuine fourth hyperscaler alongside AWS, Azure, and GCP. He also warns that organized activists — not organic local opposition — are using the same playbook that stalled nuclear construction in America to delay data-center permitting. David Sacks notes that Anthropic grew from $10B ARR on January 1 to $44B ARR by April — a trajectory he calls unlike anything Silicon Valley has ever witnessed. > *"Nobody in Silicon Valley has ever seen anything like it. Forget about the rest of the country. I mean, all we do in Silicon Valley is deal with exponentials. And still, people have never seen that kind of growth at that level of scale."* — David Sacks ## [26:48] Is Anthropic the next great monopoly? Early signals or major overreaction? David Sacks draws an extended analogy between Anthropic and John D. Rockefeller's Standard Oil, arguing that safety-first rhetoric can function as regulatory capture — building a moat that locks in the emerging duopoly of Anthropic and OpenAI while blocking competitors. He notes that if Anthropic sustains its 10× annual growth for just 18 more months it could become "the most powerful monopoly ever created in human history," dwarfing the combined Mag-7 revenue. Brad pushes back hard: Anthropic and OpenAI are still fledgling startups on a GAAP basis, Google and Amazon are producing hundreds of billions in free cash flow to fund competing models, and pre-emptive antitrust action at the starting line of AI would be "a disaster." Jason translates Brad's position as "don't mess with my paper," since Altimeter holds positions in several of these companies. Sacks clarifies his northstar is vigorous competition — but he flags Anthropic's banning of OpenClaw from using its API as a concrete anti-competitive act worth scrutiny. > *"Unless something about their current trajectory changes, Anthropic will be the most powerful monopoly ever created in human history — a trillion dollars of ARR growing at some rate. Dario calls it AGI. I call it the biggest monopoly in human history."* — David Sacks ## [35:21] "FDA for AI" freakout, how the White House thinks about AI safety Reports surfaced that the White House was considering an executive order to create an AI working group that could require pre-release safety reviews for new frontier models — triggered, according to the New York Times, by Anthropic's classified "Mythos" model reportedly alarming national-security officials. NEC Director Kevin Hassett appeared on Fox Business drawing an FDA analogy, while Treasury Secretary Scott Bessent spoke more carefully about balancing innovation and safety. Sacks calls much of it "fake news" amplified by Andrew Ross Sorkin's DealBook column, noting that Susie Wiles, the White House Chief of Staff, issued a statement walking back the FDA framing. He reveals he spoke with Hassett directly and confirms no senior official actually supports a pre-approval regime. He points to the White House's March 20 National AI Regulatory Framework as evidence the administration favors specific solutions over broad regulatory capture. The group converges on one concrete measure: KYC (Know Your Customer) requirements before frontier model API access during preview periods, plus rapid deployment of cyber-capable AI to companies like CrowdStrike and Palo Alto Networks. > *"There is a substantial faction of AI ideologues or doomers who are basically employing the classic 'never let a crisis go to waste' strategy. Yes, we do have this cyber issue that is real — everyone needs to harden their systems now. But what they're trying to do is use that issue to try and create a permanent new infrastructure in Washington."* — David Sacks ## [52:01] Flipping AI's negative perception: Giving, healthcare and education innovation Jason shifts from regulatory defense to offense: how should the tech industry proactively counter negative public perception of AI? He proposes that companies going public — Anthropic, OpenAI, SpaceX — could dedicate 1–5% of IPO proceeds to every American via "Invest America" accounts, creating tangible shared upside. He also calls for serious engagement on minimum wage and universal healthcare, arguing that a financially healthier consumer base is structurally good for capitalism itself. Brad endorses the "Invest America" concept, adding that data center host communities should receive direct benefits like free local electricity. David pivots to political salience data: AI ranks 29th out of 39 voter issues — well below cost of living and economic growth, two metrics where AI is actively deflationary and expansionary. The industry's real message should be economic delivery, not safety governance. Chamath gives tech leaders a "D-minus trending to F" for communications and calls for tangible reinvestment in America at scale. > *"I think that there's a pretty profound vibe shift with respect to tech, tech oligarchs, Silicon Valley, and particularly AI. That vibe shift has already happened on Main Street, and I think that's starting to seep into Washington."* — Chamath Palihapitiya ## [60:04] Trading the AI market, state of the economy Brad leads a comprehensive market check: AWS on a $150B run rate (28% growth), Azure at $108B (39%), Google Cloud at $80B (63%). The S&P 500 is at all-time highs, the 10-year sits at 4.3%, and inflation is under control — far better outcomes than the doom scenarios predicted around tariffs and geopolitical conflicts. S&P 500 operating margins improved from 11% in 2023 to 13% in Q1 2026, and the Mag-5's combined headcount grew only 3% over three years while revenues surged. Chamath urges caution: there is still no direct evidence AI is lifting enterprise profit margins in aggregate, and a reckoning arrives in roughly 500 days when the fork between opex reduction and revenue growth will determine whether the AI boom is real or a mirage. Jason counters that for startups the ROI is already "fait accompli" — AI-generated ad creative at Nike and DoorDash, portfolio companies shipping product at half the headcount. David credits Trump administration policies — rescinding Biden's chip-export licensing and AI-approval regime, unleashing energy permits — for creating the conditions that enabled the boom, and notes that the unemployment rate for recent college graduates has actually improved, contradicting the entry-level-job-loss narrative. > *"I think we have kind of call it 500 days where you just got to be net long. But I think it's literally in the hundreds of days from now that you're going to have to have an important reckoning moment. The people that are paying for all these tokens need to see an actual benefit."* — Chamath Palihapitiya ## Entities - **Jason Calacanis** (Person): Host and moderator; angel investor and podcast co-founder - **Chamath Palihapitiya** (Person): General partner, Social Capital; co-host; contrarian macro voice on AI ROI and market cycles - **David Sacks** (Person): Co-host; former White House AI & Crypto Czar; framed Anthropic as a potential historic monopoly using the Rockefeller analogy - **Brad Gerstner** (Person): Founder & CEO, Altimeter Capital; fifth bestie; bullish on compute stocks and AI market structure - **Dario Amodei** (Person): CEO of Anthropic; referenced as "Daario D. Rockefeller" by Sacks; party to the SpaceX compute deal - **Elon Musk** (Person): CEO of SpaceX and xAI; architect of Elon Web Services and the Colossus 1 compute lease strategy - **Anthropic** (Organization): AI lab behind Claude; grew from $10B to $44B ARR in four months; center of monopoly and FDA debates - **SpaceX / xAI** (Organization): Lessor of Colossus 1 data center to Anthropic; emerging fourth hyperscaler under EWS branding - **Elon Web Services (EWS)** (Concept): SpaceX's compute-leasing business positioned as a hyperscaler competitor to AWS, Azure, and GCP - **Mythos** (Software): Anthropic's classified cyber-capable frontier model that reportedly alarmed White House national-security officials - **KYC for AI** (Concept): Proposal to require identity verification before granting API access to frontier models during preview periods - **Invest America** (Concept): Proposal for IPO-stage tech companies to dedicate a share of proceeds to universal investment accounts for US citizens