All podcasts / All-In Podcast / Summary

All-In's Best Ideas Pitch Competition: 4 Investors Present Their Top Trades Live

2026-06-12 - 68 min - source - Read full transcript
Chamath Palihapitiya (host)Jason Calacanis (host)David Friedberg (host)Gavin BakerAaron CowenDan (Talen Energy presenter, surname garbled in source transcript - variously heard as Scherer/Draper/Dreyfus)Oleg NodelmanKyle Samani

Key insights

MGM is a hidden sum-of-parts special situation where Barry Diller's stake and buyback are the re-rating catalyst.
Aaron Cowen argued MGM trades as if it were only a Vegas casino operator, but Barry Diller has quietly built a 26% stake and bid $48/share while MGM itself bought back roughly half its float in six years. Layering in an unrecognized Osaka, Japan casino license and a speculative Dubai option, Cowen put fair value at $100-150 versus the ~$37-48 range the stock traded at when he built the pitch.
special-situations-investing
Gaming licenses aren't priced by the market until roughly three years before they open, based on the Macau precedent.
Cowen noted Osaka is the only Japanese prefecture that voted to approve a casino, and the license (approved in 2023, opening 2030) is barely mentioned in MGM's own investor materials. He drew on his 29 years covering casino stocks to note that when Wynn built out Macau, the market only started pricing the asset about three years before opening - which is roughly where MGM sits now.
special-situations-investing
AI's real bottleneck may be power, not chips - the US needs roughly 1,000x more electricity capacity than it has.
The Talen Energy presenter cited Jensen Huang's claim that AI will require a thousand times more power than currently exists, and noted the US has gone from having 2x China's power generation capacity to China now having 3x the US's. Every power source - nuclear, solar, orbital - will be needed, and the same critical-mineral supply chain (nickel superalloys, silver) is contested by data centers, power plants, and rocket launches simultaneously.
ai-power-demand
Talen Energy is a Sam Zell-style bet on buying power infrastructure below its replacement cost ahead of an AI-driven repricing.
The presenter valued Talen's ~2GW nuclear plus 6GW natural gas baseload fleet at a $25B enterprise value against a $45B replacement cost - meaning equity value alone should roughly double just to reach replacement cost, before any AI-driven demand repricing. He laid out three scenarios ($50, $70, $100+ per share of free cash flow) depending on how much new capacity and premium-priced contracting Talen adds.
ai-power-demand
Hyperscalers are locking in decades-long, above-market power deals because supply, not capital, is now the constraint.
Microsoft convinced Constellation Energy to restart the Three Mile Island nuclear reactor (site of the 1979 partial meltdown) by guaranteeing a fixed minimum price of roughly $100/MWh for 20 years, well above the prevailing ~$50/MWh spot price - illustrating how desperate data-center operators have become to secure firm baseload power.
ai-power-demand
Radiopharmaceuticals can be de-risked earlier in clinical trials than most biotech because doctors can image whether the drug actually reaches the tumor.
Oleg Nodelman explained that Actus Oncology's platform attaches radioactive payloads to mini-proteins; because the payload is detectable via imaging, physicians can directly verify target engagement in early trials rather than inferring efficacy indirectly, which significantly reduces development risk relative to typical small-molecule or antibody programs.
biotech-radiopharma
Radiopharma has a moat against Chinese drug replication that most of biotech lacks, because the isotopes themselves are scarce.
Nodelman noted China has aggressively reverse-engineered US-patented biologics (enabled by loopholes the Amgen v. Sanofi Supreme Court ruling exposed in composition-of-matter patents), but Actus's actinium isotope is a byproduct of Cold War-era US radium-233 nuclear programs and isn't available through China's different (enriched uranium/plutonium) nuclear supply chain, giving the modality real competitive protection.
biotech-radiopharma
GeoDNET out-built three legacy RTK network operators combined in about three years by crowdsourcing infrastructure with token payments.
Kyle Samani explained that Trimble, Hexagon, and TopCon spent 20-30 years building roughly 12,000 combined RTK base stations; GeoDNET, founded in 2021, now has roughly 22,000 nodes by paying ordinary individuals a few hundred dollars in GEO tokens to host a base station on their roof, turning a capital-intensive telecom-style buildout into a crowdsourced, CapEx-light network now live in 150 countries.
crypto-infrastructure-tokenomics
GeoDNET's token functions as a de facto revenue-share security, with 80% of network revenue used for open-market token buybacks.
Samani described GeoDNET's ~$11M annualized revenue (growing 3x year-over-year) as being split so 80% funds open-market GEO token purchases (visible on-chain) and 20% covers R&D and business development, meaning token holders directly capture the majority of revenue growth - a structure whose legal treatment depends on frameworks like the pending Clarity Act.
crypto-infrastructure-tokenomics
Position sizing should weigh liquidity and downside protection as much as conviction, not just upside magnitude.
Chamath said he loved all four pitches equally as ideas but would size them very differently: he could put tens of millions into MGM or Talen without moving the market, but could barely deploy a million dollars into GeoDNET's token before moving its price, and treated the biotech pitch as effectively binary (zero or 10x) given its illiquidity and single-catalyst structure.
investment-risk-frameworks
Guest judge Gavin Baker split the four pitches into 'lottery tickets' versus assets with hard downside protection.
Baker grouped Actus (biotech) and GeoDNET (crypto) together as high-variance bets with a real chance of going to zero if the science or thesis doesn't pan out, versus MGM and Talen, which he viewed as having tangible asset-backed downside protection (real estate and a firm takeover bid for MGM; cash-generative power plants for Talen) even in a bear case.
investment-risk-frameworks

Companies

Techniques and frameworks

Summary

This live event, modeled explicitly on the Sohn Investment Conference (which Chamath describes recounting his own history there, including being laughed at by David Einhorn in 2015 for predicting Amazon would become a trillion-dollar company), gathers four professional investors to each pitch one high-conviction "best idea" trade to the All-In hosts (Chamath Palihapitiya, Jason Calacanis, and David Friedberg - David Sacks is noted as absent) plus guest judge Gavin Baker. Each presenter gets roughly six minutes followed by audience and judge questions, and the event ends with both an audience vote and a judges' ranking.

Aaron Cowen of Suvretta Capital opens with MGM Resorts, arguing the market is mispricing it purely as a Las Vegas operator while ignoring that activist Barry Diller has quietly built a 26% stake and bid $48/share, that MGM itself has bought back roughly half its float, and that the company holds an unrecognized casino license in Osaka, Japan (the only Japanese prefecture approved for gaming) plus a speculative option on a future Dubai casino. Drawing on his 29 years in hedge funds, Cowen argues gaming licenses historically aren't priced by the market until roughly three years before opening, per the Wynn Macau precedent, putting MGM's fair value at $100-150 versus its then-current price in the high $30s to high $40s.

The second presenter (introduced only as "Dan," with the transcript's auto-generated captions rendering his surname inconsistently as Scherer, Draper, and Dreyfus in different spots) pitches Talen Energy as a bet on the coming AI power supercycle, applying Sam Zell's framework of buying hard assets below replacement cost. He values Talen's nuclear and natural-gas power fleet at a steep discount to its $45B replacement cost, cites Jensen Huang's claim that AI will need roughly 1,000x more power than currently exists, and points to Microsoft's 20-year, above-market power deal to restart the Three Mile Island reactor as evidence that hyperscalers are now supply-constrained, not capital-constrained, in securing electricity.

Oleg Nodelman of EcoR1 Capital pitches Actus Oncology (ticker AKTS), a radiopharmaceutical company whose mini-protein-guided isotope payloads can be verified reaching tumors via imaging - a rare de-risking advantage in biotech. He frames the pitch through an extended war-on-cancer analogy (surgery as medieval siege, chemo as indiscriminate mustard-gas-derived poison, radiopharma as precision microdrones) and argues the modality is protected from Chinese drug replication because its actinium isotope is a scarce byproduct of Cold War-era US nuclear programs. Kyle Samani, formerly of Multicoin Capital, closes with GeoDNET, a decentralized RTK positioning network (ticker GEOD) that crowdsourced its way past legacy operators Trimble, Hexagon, and TopCon combined by paying ordinary people in tokens to host base stations, and now routes 80% of its revenue into open-market token buybacks for customers like John Deere, DJI, and TomTom.

In judging, Gavin Baker splits the four ideas into "lottery ticket" bets (Actus and GeoDNET, which carry real odds of going to zero) versus assets with tangible downside protection (MGM's real estate and Diller's bid floor, Talen's cash-generative power plants), while flagging AI regulatory risk as Talen's main external threat. Chamath echoes this framing but through a liquidity lens, noting he could deploy tens of millions into MGM or Talen without moving markets but barely a million dollars into GeoDNET's thin token liquidity - illustrating that position sizing depends on downside protection and liquidity as much as raw conviction. The event closes with MGM winning both the audience and judges' vote.

Notable Quotes

"We do not need AI demand to keep the power markets incredibly tight for the next 20 years. AI demand just turbocharges. That's all it does." - Dan (Talen Energy presenter)

"We're one of the few funds not managed by PhDs or MDs, and that's by design, because we don't want to fall in love with the science. We fall in love with the risk reward." - Oleg Nodelman

"You have Barry Diller, who's the legend, aggressively buying the stock, and it's also now 80% of his NAV." - Aaron Cowen

"This thing looks like a natural telecom... Telecoms naturally form monopolies historically. I think the same is likely to be true here." - Kyle Samani

"I feel about crypto exactly the way I do about snowboarding... I've got 25 years of lessons, learnings, pain, scars from investing in equities and public securities." - Gavin Baker