{
 "episode_id": "all-in-2026-06-06-the-ipo-comeback-why-tech-giants-are-finally-going-public-al",
 "title": "The IPO Comeback: Why Tech Giants Are Finally Going Public | All-In Liquidity IPO Panel",
 "podcast": "all-in",
 "published_at": "2026-06-06",
 "speakers": [
  "Jason Calacanis",
  "Chamath Palihapitiya",
  "Andrew Feldman",
  "Will Marshall",
  "Brad Gerstner",
  "David Sacks",
  "Friedberg"
 ],
 "summary": "An All-In live panel on the IPO comeback, moderated alongside Cerebras CEO Andrew Feldman and Planet Labs CEO Will Marshall, with David Sacks driving the liquidity discussion. Feldman and Marshall reflect on what going public actually changes (mostly credibility and capital, not the day-to-day work), and Sacks argues the decade-long 'stay private forever' era pushed by Andreessen is reversing, with more portfolio companies now eyeing IPOs at $1-5B so public-market investors can capture the upside. Marshall lays out Planet's Earth-imaging business and a bull case for space-based data centers becoming cheaper than terrestrial ones once launch costs fall to ~$200-300/kg, predicting most compute moves to space within a decade. Feldman explains Cerebras's contrarian bet on a dinner-plate-sized chip with memory next to compute, claiming 15-18x inference speedups over GPUs and arguing a GPU-shaped architecture can never beat NVIDIA. Feldman pushes back on the space-compute timeline, citing unsolved clustering and the 'last 10%' problem, while agreeing it will eventually happen.",
 "topics": [
  {
   "name": "Intro montage and Davos anecdote",
   "timestamp": "0:00",
   "summary": "The panelists and guests are introduced over a highlight montage, followed by Jason's story about being pulled into Davos to moderate."
  },
  {
   "name": "What actually changes when you go public",
   "timestamp": "2:42",
   "summary": "Feldman and Marshall describe the IPO experience — lots of process and ceremony, but the core business and engineering work continues unchanged; the main gains are capital, profile, and credibility with customers."
  },
  {
   "name": "Cerebras's IPO path and pricing",
   "timestamp": "4:39",
   "summary": "Sacks recounts Cerebras's hard decade-long road to a public listing, including UAE-investor scrutiny under the prior administration, and the eventual IPO pricing and market-cap outcome."
  },
  {
   "name": "Planet Labs business, military mix, and satellite miniaturization",
   "timestamp": "8:22",
   "summary": "Marshall explains Planet's daily Earth-imaging fleet, its customer mix (about 60% military/security), and the launch-cost and satellite-miniaturization trends enabling the business."
  },
  {
   "name": "Space-based data centers",
   "timestamp": "13:16",
   "summary": "Marshall makes the case that space data centers become cheaper than terrestrial ones as launch costs fall and solar power is continuous in orbit; Feldman counters that clustering and the 'last 10%' remain unsolved and the timeline is longer."
  },
  {
   "name": "The rebirth of silicon and Cerebras's architecture bet",
   "timestamp": "19:42",
   "summary": "Feldman frames AI as opening compute to images and language, and explains Cerebras's two bets — dedicated silicon and a non-GPU architecture using a huge chip with memory beside compute — claiming large inference speedups."
  },
  {
   "name": "Liquidity, lockups, and capturing upside in public markets",
   "timestamp": "24:47",
   "summary": "Sacks, Feldman, and Marshall debate whether early investors should hold post-IPO; they argue more value is often created publicly, cite Planet's 10x public-market move and a Mongo counter-example, and discuss the innovative 'dribble lockup.'"
  },
  {
   "name": "The pendulum swinging back toward earlier IPOs",
   "timestamp": "29:31",
   "summary": "Sacks argues the 'stay private forever' era is reversing and portfolio companies want to go public sooner; Friedberg adds that public-market scrutiny sharpens execution, and Marshall closes on AI-plus-space 'planetary intelligence.'"
  }
 ],
 "entities": [
  {
   "name": "Cerebras",
   "type": "company",
   "salience": "central",
   "role_in_episode": "Recently-IPO'd AI chip company whose CEO is on the panel; case study for going public and for non-GPU silicon."
  },
  {
   "name": "Planet Labs",
   "type": "company",
   "salience": "central",
   "role_in_episode": "Earth-imaging satellite company whose CEO is on the panel; case study for public-market value creation and space data centers."
  },
  {
   "name": "Andrew Feldman",
   "type": "person",
   "salience": "central",
   "role_in_episode": "Founder and CEO of Cerebras; explains the IPO experience, Cerebras's architecture, and pushes back on the space-compute timeline."
  },
  {
   "name": "Will Marshall",
   "type": "person",
   "salience": "central",
   "role_in_episode": "Co-founder and CEO of Planet Labs; explains the Earth-imaging business and the bull case for space-based data centers."
  },
  {
   "name": "David Sacks",
   "type": "person",
   "salience": "central",
   "role_in_episode": "All-In host and investor; drives the liquidity/IPO discussion and argues the private-markets pendulum is swinging back."
  },
  {
   "name": "NVIDIA",
   "type": "company",
   "salience": "discussed",
   "role_in_episode": "Incumbent GPU maker used as the benchmark Cerebras's architecture is designed to beat."
  },
  {
   "name": "SpaceX",
   "type": "company",
   "salience": "discussed",
   "role_in_episode": "Referenced as an upcoming mega-IPO whose lockup and valuation dynamics are contrasted with smaller public companies."
  },
  {
   "name": "IPOs",
   "type": "concept",
   "salience": "central",
   "role_in_episode": "The central topic — the comeback of going public and the timing, value-capture, and lockup dynamics around it."
  },
  {
   "name": "space-based data centers",
   "type": "concept",
   "salience": "central",
   "role_in_episode": "A major debated trend — putting compute in orbit; Marshall is bullish, Feldman more cautious on timeline."
  },
  {
   "name": "OpenAI",
   "type": "company",
   "salience": "mentioned",
   "role_in_episode": "Cited as a Cerebras inference customer benefiting from faster-than-GPU performance, and as part of the private-mega-cap cohort."
  },
  {
   "name": "Google",
   "type": "company",
   "salience": "mentioned",
   "role_in_episode": "Planet's partner and shareholder; collaborator on the space-data-center cost study and on launching TPUs into space."
  },
  {
   "name": "Altimeter Capital",
   "type": "company",
   "salience": "mentioned",
   "role_in_episode": "Brad Gerstner's investment firm; a Cerebras investor under lockup discussed in the distribution debate."
  },
  {
   "name": "Andreessen Horowitz",
   "type": "company",
   "salience": "mentioned",
   "role_in_episode": "Cited as the firm that pushed the 'stay private forever' thesis Sacks says is now reversing."
  },
  {
   "name": "Starship",
   "type": "product",
   "salience": "mentioned",
   "role_in_episode": "SpaceX launch vehicle cited as the trajectory that will bring launch costs down enough for space data centers."
  }
 ],
 "claims": [
  {
   "holder": "Andrew Feldman",
   "claim": "Going public involves an enormous amount of wasteful process and ceremony while changing nothing about the core parts of the business.",
   "stance": "opinion",
   "conviction": 0.75,
   "subjects": [
    "IPOs",
    "Cerebras"
   ],
   "timestamp": "3:03",
   "quote": "it's really difficult to overestimate the amount of garbage that's involved in going public"
  },
  {
   "holder": "Andrew Feldman",
   "claim": "After an IPO, the important parts of the business are unchanged — only the cash position and constituents shift.",
   "stance": "assertion",
   "conviction": 0.8,
   "subjects": [
    "IPOs",
    "Cerebras"
   ],
   "timestamp": "3:03",
   "quote": "not a damn thing changes in the important parts of your business"
  },
  {
   "holder": "David Sacks",
   "claim": "Cerebras's path to going public was exceptionally hard — nine and a half years of difficulty followed by twelve months of everything becoming easy.",
   "stance": "assertion",
   "conviction": 0.85,
   "subjects": [
    "Cerebras"
   ],
   "timestamp": "4:39",
   "quote": "everything was really hard until it got really easy"
  },
  {
   "holder": "Will Marshall",
   "claim": "Being a public company gives Planet credibility with major customers who need assurance the company will be around long-term.",
   "stance": "opinion",
   "conviction": 0.8,
   "subjects": [
    "Planet Labs",
    "IPOs"
   ],
   "timestamp": "8:27",
   "quote": "the maturing event gives you more credibility to various customers"
  },
  {
   "holder": "Will Marshall",
   "claim": "AI models are only as valuable as the underlying data, so Planet's Earth-imaging data is what makes AI applications possible.",
   "stance": "assertion",
   "conviction": 0.75,
   "subjects": [
    "Planet Labs",
    "space-based data centers"
   ],
   "timestamp": "10:33",
   "quote": "AI is only as good as the data it's trained upon"
  },
  {
   "holder": "Will Marshall",
   "claim": "Once launch costs fall to roughly $200-300 per kilogram, it will simply be cheaper to put data centers in space than on the ground.",
   "stance": "assertion",
   "conviction": 0.8,
   "subjects": [
    "space-based data centers",
    "Planet Labs"
   ],
   "timestamp": "13:16",
   "quote": "when launch costs come down to about 200 to 300 dollars a kilogram, it would be cheaper, just simply cheaper, to put the data centers in space"
  },
  {
   "holder": "Will Marshall",
   "claim": "Satellite miniaturization — shrinking from billion-dollar 20-ton satellites to a few kilograms — is an underappreciated driver as important as falling launch costs.",
   "stance": "opinion",
   "conviction": 0.75,
   "subjects": [
    "Planet Labs",
    "space-based data centers"
   ],
   "timestamp": "12:10",
   "quote": "a thing that people don't know that is actually perhaps more important, is that we've had a miniaturization of satellites"
  },
  {
   "holder": "Andrew Feldman",
   "claim": "Building clusters in space for communication between chips is an unsolved hard problem — we are not even good at it on the ground.",
   "stance": "assertion",
   "conviction": 0.75,
   "subjects": [
    "space-based data centers"
   ],
   "timestamp": "18:08",
   "quote": "I think we're not super good yet at building the clusters in space necessary for the communication"
  },
  {
   "holder": "Andrew Feldman",
   "claim": "Space-based compute will certainly happen, but on a longer time frame than Planet implies, and the risk is a self-driving-style 'last 10% takes 80% of the time' problem.",
   "stance": "prediction",
   "conviction": 0.7,
   "subjects": [
    "space-based data centers"
   ],
   "timestamp": "18:34",
   "quote": "I've got it in a slightly different time frame, but one that certainly will occur"
  },
  {
   "holder": "Andrew Feldman",
   "claim": "If you build a GPU-shaped architecture, your odds of beating NVIDIA are approximately zero, because they have eaten all the low-hanging fruit.",
   "stance": "assertion",
   "conviction": 0.85,
   "subjects": [
    "Cerebras",
    "NVIDIA"
   ],
   "timestamp": "19:42",
   "quote": "if you build a GPU, the odds that you're better than NVIDIA in our view are approximately zero"
  },
  {
   "holder": "Andrew Feldman",
   "claim": "To be 20x better than an incumbent, your architecture fundamentally cannot resemble theirs.",
   "stance": "assertion",
   "conviction": 0.85,
   "subjects": [
    "Cerebras",
    "NVIDIA"
   ],
   "timestamp": "19:42",
   "quote": "if you want to be 20 times better than somebody, your architecture can't look like them"
  },
  {
   "holder": "Andrew Feldman",
   "claim": "Latency is decisive for AI — there is no market for slow inference, so AI must be delivered in real time.",
   "stance": "opinion",
   "conviction": 0.8,
   "subjects": [
    "Cerebras",
    "OpenAI"
   ],
   "timestamp": "19:42",
   "quote": "You will not wait for AI. We have to deliver it to you in real time"
  },
  {
   "holder": "Andrew Feldman",
   "claim": "Historically more money is made after an IPO than before, in both percentage and absolute terms, so investors should hold post-lockup.",
   "stance": "assertion",
   "conviction": 0.85,
   "subjects": [
    "IPOs",
    "Altimeter Capital"
   ],
   "timestamp": "27:39",
   "quote": "I think historically more money is made after IPO than before"
  },
  {
   "holder": "Will Marshall",
   "claim": "Most of Planet's early investors should hold their shares even longer because most of the value was created in the public markets, years after the listing.",
   "stance": "opinion",
   "conviction": 0.7,
   "subjects": [
    "Planet Labs",
    "IPOs"
   ],
   "timestamp": "25:38",
   "quote": "which is really smart on their part, obviously, I think they should hold on even more"
  },
  {
   "holder": "David Sacks",
   "claim": "Had Cerebras gone public ~18 months earlier at around $10B instead of $50B, the subsequent 5x would have accrued to public-market investors instead of private ones.",
   "stance": "assertion",
   "conviction": 0.7,
   "subjects": [
    "Cerebras",
    "IPOs"
   ],
   "timestamp": "28:52",
   "quote": "he would have been public 18 months ago, probably at $10 billion rather than $50 billion"
  },
  {
   "holder": "David Sacks",
   "claim": "The public markets are shifting back toward earlier IPOs, reversing the decade in which Andreessen pushed companies to stay private forever.",
   "stance": "prediction",
   "conviction": 0.7,
   "subjects": [
    "IPOs",
    "Andreessen Horowitz"
   ],
   "timestamp": "29:31",
   "quote": "I actually think the public markets may be shifting back in this direction"
  },
  {
   "holder": "David Sacks",
   "claim": "The private-markets pendulum is swinging back, with portfolio companies now wanting to follow Planet Labs and go public.",
   "stance": "opinion",
   "conviction": 0.75,
   "subjects": [
    "IPOs"
   ],
   "timestamp": "29:31",
   "quote": "And I see the pendulum swinging back"
  },
  {
   "holder": "Friedberg",
   "claim": "Going public sooner and submitting to public-market scrutiny sharpens a company's focus and makes innovation better.",
   "stance": "opinion",
   "conviction": 0.75,
   "subjects": [
    "IPOs"
   ],
   "timestamp": "30:08",
   "quote": "getting public sooner, having the scrutiny of public markets, having the scrutiny of having to deliver, sharpens the focus"
  },
  {
   "holder": "Will Marshall",
   "claim": "Today's LLMs are 'blind' to the real world because they're trained only on internet text; feeding them real-world Earth data will unlock 'planetary intelligence' and a huge new economy.",
   "stance": "opinion",
   "conviction": 0.75,
   "subjects": [
    "Planet Labs",
    "space-based data centers"
   ],
   "timestamp": "31:11",
   "quote": "But they don't know shit about the real world. I call them blind"
  }
 ],
 "predictions": [
  {
   "holder": "Will Marshall",
   "prediction": "On the current Starship trajectory, launch costs will fall enough (toward the ~$200-300/kg threshold) within two to three years to make space data centers cheaper than terrestrial ones.",
   "timeframe": "two to three years",
   "subjects": [
    "space-based data centers",
    "Starship",
    "Planet Labs"
   ],
   "timestamp": "13:16",
   "quote": "On the current trajectory with Starship in particular, I would expect the launch costs to come down there in two to three years",
   "due": "2029-06-06"
  },
  {
   "holder": "Will Marshall",
   "prediction": "Within 10 years, most compute will be placed in space, becoming a multi-trillion-dollar market bigger than any existing space business.",
   "timeframe": "within 10 years",
   "subjects": [
    "space-based data centers",
    "Planet Labs"
   ],
   "timestamp": "13:16",
   "quote": "I think no question within 10 years, most compute will be put in space, which to give you a sense is a lot of money, like trillions",
   "due": "2036-06-06"
  },
  {
   "holder": "David Sacks",
   "prediction": "Many companies in the panelists' portfolios will go public at $1-5B valuations as the IPO market reopens.",
   "timeframe": "unspecified",
   "subjects": [
    "IPOs"
   ],
   "timestamp": "29:31",
   "quote": "a lot of the companies in our portfolios are now thinking about going public at a billion or three billion or five billion"
  },
  {
   "holder": "Andrew Feldman",
   "prediction": "Space-based compute will eventually happen, though on a longer time frame than Planet projects.",
   "timeframe": "unspecified (longer than Planet's)",
   "subjects": [
    "space-based data centers"
   ],
   "timestamp": "18:34",
   "quote": "I've got it in a slightly different time frame, but one that certainly will occur"
  }
 ],
 "facts": [
  {
   "statement": "Cerebras priced its IPO at $185, with the range taken up twice.",
   "subjects": [
    "Cerebras"
   ],
   "value": "$185 IPO price",
   "timestamp": "4:39",
   "quote": "They priced the IPO at 185, which was up, the range was taken up two times",
   "source": "David Sacks"
  },
  {
   "statement": "Cerebras stock opened around $320, trades around $230, for roughly $50-60B market cap.",
   "subjects": [
    "Cerebras"
   ],
   "value": "~$320 open, ~$230 now, $50-60B cap",
   "timestamp": "4:39",
   "quote": "the stock opened at $320 a share, I think. Today it's at 230 bucks a share, 50, $60 billion in market cap",
   "source": "David Sacks"
  },
  {
   "statement": "One of Cerebras's investors was the UAE, which created scrutiny under the prior administration.",
   "subjects": [
    "Cerebras"
   ],
   "timestamp": "4:39",
   "quote": "One of their investors was the UAE",
   "source": "David Sacks"
  },
  {
   "statement": "Planet Labs stock rose from $5 to $50 a share over 12 months, a 10x move.",
   "subjects": [
    "Planet Labs"
   ],
   "value": "$5 -> $50, ~10x",
   "timestamp": "7:59",
   "quote": "your stock has gone from 5 bucks a share to 50 bucks a share, some 10x move in the public markets",
   "source": "David Sacks"
  },
  {
   "statement": "About 60% of Planet Labs' revenue today is military/security.",
   "subjects": [
    "Planet Labs"
   ],
   "value": "~60% military",
   "timestamp": "11:32",
   "quote": "It's about 60% of our revenue today",
   "source": "Will Marshall"
  },
  {
   "statement": "Planet operates the largest Earth-imaging fleet, about 200 satellites, imaging the entire Earth every day.",
   "subjects": [
    "Planet Labs"
   ],
   "value": "~200 satellites, daily full-Earth imaging",
   "timestamp": "10:33",
   "quote": "We have the largest Earth imaging fleet, about 200 satellites, they image the entire Earth every day",
   "source": "Will Marshall"
  },
  {
   "statement": "Rocket/launch costs have come down roughly 4-5x over the last 10 years.",
   "subjects": [
    "space-based data centers"
   ],
   "value": "~4-5x lower over 10 years",
   "timestamp": "12:10",
   "quote": "rocket costs have come down about 4 or 5x over the last 10 years",
   "source": "Will Marshall"
  },
  {
   "statement": "Marshall estimates a $75-100B market for Earth-observation data.",
   "subjects": [
    "Planet Labs"
   ],
   "value": "$75-100B market",
   "timestamp": "13:16",
   "quote": "I estimate there's a $75 to $100 billion market just on Earth observation",
   "source": "Will Marshall"
  },
  {
   "statement": "Current launch cost is about $1,000/kg, down roughly 10x over the last 10 years.",
   "subjects": [
    "space-based data centers"
   ],
   "value": "~$1,000/kg, ~10x lower over 10 years",
   "timestamp": "13:16",
   "quote": "we're about a thousand dollars a kilogram just over that in right today, but that's come down about 10x in the last 10 years",
   "source": "Will Marshall"
  },
  {
   "statement": "Planet is partnering with Google to launch Google's TPUs into space and has already launched NVIDIA GPUs into space.",
   "subjects": [
    "Planet Labs",
    "Google",
    "NVIDIA"
   ],
   "timestamp": "13:16",
   "quote": "we're partnering with Google to launch some of their TPUs into space",
   "source": "Will Marshall"
  },
  {
   "statement": "When OpenAI runs inference on Cerebras, it is 15-18x faster than a GPU.",
   "subjects": [
    "Cerebras",
    "OpenAI"
   ],
   "value": "15-18x faster than a GPU",
   "timestamp": "19:42",
   "quote": "when OpenAI uses us, we're 15 or 18 times faster than a GPU",
   "source": "Andrew Feldman"
  },
  {
   "statement": "Planet Labs went public in 2021 at a $2B valuation via a SPAC.",
   "subjects": [
    "Planet Labs"
   ],
   "value": "2021, $2B, SPAC",
   "timestamp": "25:01",
   "quote": "Okay. So your investors come in, you go public at 2 billion via a SPAC",
   "source": "David Sacks"
  },
  {
   "statement": "Sacks's firm invested in Mongo pre-IPO at ~$1B and distributed shares at $3-4B; it then rose to $50B over the next 24 months.",
   "subjects": [
    "IPOs"
   ],
   "value": "$1B -> $3-4B distribution -> $50B in 24 months",
   "timestamp": "26:15",
   "quote": "We distributed the shares, I think, at three or four billion, and then it went to 50 billion over the course of the next 24 months",
   "source": "David Sacks"
  },
  {
   "statement": "Cerebras innovated a 'dribble lockup' with its banks, releasing shares over six months against performance hurdles, which SpaceX is expected to mirror.",
   "subjects": [
    "Cerebras",
    "SpaceX"
   ],
   "timestamp": "26:15",
   "quote": "over six months, the shares can be dribbled out according to a bunch of performance hurdles, which SpaceX is going to have a very similar lockup to you",
   "source": "David Sacks"
  }
 ],
 "notable_quotes": [
  {
   "speaker": "Andrew Feldman",
   "timestamp": "6:24",
   "quote": "I think the answer is by getting it wrong for a decade. That's really the right way to get timing right.",
   "why": "A memorable, counterintuitive line on how Cerebras 'got the IPO timing right' — by failing for ten years first."
  },
  {
   "speaker": "Will Marshall",
   "timestamp": "0:20",
   "quote": "We are indexing the earth and making it searchable.",
   "why": "Crisp framing of Planet's entire mission as a Google-for-the-physical-world analogy."
  },
  {
   "speaker": "Andrew Feldman",
   "timestamp": "19:42",
   "quote": "How big is the market for slow search today?",
   "why": "A vivid rhetorical argument that latency is existential for AI products."
  },
  {
   "speaker": "Will Marshall",
   "timestamp": "31:11",
   "quote": "instead of having large language models, large Earth models, or instead of AI, planetary intelligence",
   "why": "Coins the panel's big-idea framing for fusing satellite data with AI."
  },
  {
   "speaker": "David Sacks",
   "timestamp": "29:31",
   "quote": "And I see the pendulum swinging back.",
   "why": "The thesis statement of the episode — the era of staying private forever is ending."
  },
  {
   "speaker": "Andrew Feldman",
   "timestamp": "4:39",
   "quote": "not a damn thing changes in the important parts of your business",
   "why": "A blunt, quotable reality-check on what an IPO does and doesn't change for a company."
  }
 ],
 "extracted_by": "agent",
 "prompt_version": "episode-extract@3",
 "_file": "all-in-2026-06-06-the-ipo-comeback-why-tech-giants-are-finally-going-public-al.json"
}