Epstein Files Flop, State of the Market, Autonomous Robots, Trump's Gold Card, Friedberg on Jeopardy
Speakers
Chamath Palihapitiya
— CEO, Social Capital
Jason Calacanis
— Angel Investor / Founder, LAUNCH
David Friedberg
— CEO, The Production Board
Summary
Jason Calacanis, Chamath Palihapitiya and David Friedberg (David Sacks is absent) open on the underwhelming first Epstein files release and Friedberg's Celebrity Jeopardy win, then move to markets. They debate humanoid robots, drones and autonomous vehicles after Figure pulled its home beta forward, compare the Stripe and Adyen annual reports, and dig into global stablecoin infrastructure and how much faster AI companies reach revenue than traditional SaaS. The market segment covers slowing growth from tariffs, immigration curbs and DOGE austerity, Mag 7 valuations priced to perfection, a 10-year yield that has fallen half a point in a month, and Chamath's great reset theory that cementing political power requires walking down stock and real-estate prices. They close on Trump's $5 million gold card, whether retail should be allowed into private markets, a USPS restructuring and Bezos narrowing the Washington Post opinion page.
- Figure pulled its in-home robot beta forward; the hosts debate actuator dexterity, model generality and the two-robot grocery-sorting demo.
- Friedberg widens the robot category to drones and autonomous vehicles but expects single-purpose machines to commercialize before general-purpose humanoids.
- Chamath's Stripe report takeaways: an underappreciated product ecosystem, global stablecoin infrastructure, and AI companies hitting $5M ARR in 24 months versus 37 for SaaS.
- Legacy enterprise software renewal cycles, Salesforce included, are getting harder to justify as AI replacements appear, though regulated industries still cannot tolerate hallucinations.
- Chamath is cautious on equities: tariffs, slower immigration and DOGE austerity slow growth to about 1.5%, and the Mag 7 is priced to perfection.
- The 10-year has fallen from 5% pre-election and 4.78% in January to 4.26%; Chamath sees sub-4% possible with $10 trillion to refinance, Friedberg reads it as contained inflation.
- Great reset theory: a durable coalition of asset-light voters implies power is cemented by walking down stocks and real estate.
- Friedberg says ordinary investors should own the S&P 500 rather than startups; Chamath counters that the cap-weighted index is now effectively the S&P 7.
Ideas
Humanoid robots still years from real usefulness
Chamath argues the humanoid robot category is bounded by two problems: the generalized AI model is not good enough yet (Figure just publicly cancelled its OpenAI deal and announced its own model of unclear provenance), and the actuators are good but not great, so physical dexterity is still relatively limited. That keeps these robots from being super functional for the next couple of years, even though the demo of two robots communicating semantically to sort a bag of groceries unsupervised is a real breakthrough. Once the dexterity problem is solved, a robot that can sort groceries, make food, do laundry and mow the lawn becomes genuinely useful.
Robotics is the underappreciated investment category
Jason says robots are the category people are sleeping on. He points to immediate mundane demand (ranch work such as weed whacking, trimming hedges, hauling wood and collecting chicken eggs, where a robot working 24 hours a day can afford to be slow) and to the fact that single-purpose machines are already cheap and commercially deployed: robot lawnmowers at about $1,000 that he already sees on lawns in Austin, and Roombas at $300-400. He notes the industry has called it the year of robots for 30 years, but with home beta testing now being pulled forward it feels like this is finally it.
Automation accelerating; single-purpose devices arrive first
Friedberg sticks with his prediction that this is the year of robots and defines the category broadly: dexterous humanoid automation, drones and autonomous vehicles all combine a mechanical response to a machine-vision system, and that combination has accelerated this year. He adds a qualifier that shapes the order of adoption: a general-purpose humanoid is an ambitious and technically very hard roadmap, so vertically integrated, utility-specific machines that do exactly one thing (deliver something by air, drive food to you, load and unload dishes) get commercialized before general-purpose systems do.
Stablecoin infrastructure adoption is going global
Chamath repeats a view he has held for a while: the rise of stablecoins and of global stablecoin infrastructure is the most interesting structural development in payments, with national governments in India and Brazil, and slowly the United States, moving to embrace them. He frames the rails as ultimately ledger entries between two systems of record, which is why the winning approach is to facilitate payments on top of what already exists rather than launch yet another coin. The episode notes Tether at roughly $143B outstanding and USDC at roughly $56B after only a few years.
AI companies scale revenue far faster
Chamath's third takeaway from Stripe's annual report is that AI companies monetize at a speed he has never seen before: the average SaaS company took 37 months to reach $5M of annualized revenue, while by 2024 the top 100 AI companies got there in 24 months, and his own 8090 reached $5M of revenue in three months. The selling motion is different because the ROI and the cost savings versus traditional enterprise software are obvious, so the sale is more straightforward, the revenue is bigger and it happens faster. He treats the Stripe data as validation that growth in this industry is unlike anything he has seen.
Legacy enterprise software renewals getting harder
Chamath says his 8090 customers are not yet driven by urgency to adopt AI but by frustration with what he calls the software industrial complex: renewal cycles for large incumbents such as Salesforce are getting harder and harder to justify, so buyers are now willing to take bets on alternatives. The real opportunity is a repeatable pattern that replaces that big legacy software spend, which implies sustained pressure on incumbent enterprise software vendors. He notes the limiting factor: in regulated environments such as healthcare, finance, real estate, power and aerospace, hallucinations carry consequences that are not solved yet, so displacement is easiest where errors are self-evident, like code.
Growth slowing; best equity gains already had
Chamath's great reset working theory: the coalition that cements durable political power combines the asset-light working and middle class who own neither homes nor stock with patriotic business owners and technologists, and that bloc is growing faster than the college-educated, $100k-plus cohort that votes the other way. Feeding those constituents is bad news for the stock market and for asset owners because rewarding asset holders does not reward the voters who decide elections. His conclusion is that cementing power over multiple elections requires walking down asset markets in a meaningful way, both stocks and real estate, and he later repeats that if accreditation and access rules never change, debasing assets is the only remaining path to broader ownership. He flags it explicitly as a theory he could change as he gets more data.
Mag 7 priced to perfection
Looking at forward P/E charts, Chamath notes the Mag 7 premium is compressing toward everybody else as the rest of the market captures back ground and investors reprocess how much real upside the mega-caps still have. The Mag 7 is priced to perfection, so owning it requires believing the world stays exactly the way it is; otherwise there is some amount of mean reversion ahead. On the margin he therefore thinks the stock market is a little expensive and not particularly attractive.
Ten-year yield could fall below 4%
Chamath reads the meaningful compression in the 10-year yield as the bond market deciding to give the administration credit that DOGE and tariffs will work. That matters because roughly $10 trillion of debt has to be refinanced in the next six months, so lower yields are very good for Bessent and Trump, and with a good string of data he thinks the 10-year could even get under 4%. The UK austerity precedent supports the mechanism: while the government pursued deficit reduction, the bond market kept rates relatively low.
Bond market signals contained inflation ahead
Friedberg tracks the 10-year from its 5% peak two weeks before the election, to a second peak of 4.78% in the second week of January, to 4.26% at taping: a full half point in a month. He reads that as the bond market telling you that inflation and growth expectations for the next decade are contained, and that it is a reasonable sign there will be no rampant inflation from the policy mix being pursued. Despite very wide uncertainty around tariffs, tax cuts and spending cuts, his decoded signal is that the net effect is generally somewhat deflationary rather than inflationary.
Ordinary investors should just own S&P
Friedberg takes the opposite side of Jason's accreditation argument and says ordinary investors should simply buy the S&P 500 index: proven, scaled, audited, profitable, well-vetted public companies with boards and fiduciary responsibility. His reasoning is that there is no shortage of startups, only a shortage of good ones; flooding the private market with capital funds companies that should not be funded and eats people's money; and unsophisticated entrants into any new market face adverse selection, with predatory pitches and fancy PowerPoints taking their money. He adds that even the best Silicon Valley venture firms, with the most sophisticated people, were not able to beat the Nasdaq over years.
Up Next
This All-In Podcast video, published March 01, 2025, features Chamath Palihapitiya, Jason Calacanis, David Friedberg discussing Humanoid robots, ROBO, Autonomous vehicles, STABLECOINS, AI software, CRM, SPY, US Real Estate, MAGS, IEF. 11 trade ideas extracted by AI with direction and confidence scoring.
Speakers: Chamath Palihapitiya, Jason Calacanis, David Friedberg · Tickers: Humanoid robots, ROBO, Autonomous vehicles, STABLECOINS, AI software, CRM, SPY, US Real Estate, MAGS, IEF