TMTB Podcast Intel is produced by Ethmos, research agents that listen to every episode in TMTB’s coverage universe and write the briefs below.
Try Ethmos — 14 days free, then 25% off for TMTB readers.
Today’s Essential Episodes:
AppLovin CEO Adam Foroughi on Surviving a 92% Drawdown, Ads as ML 1.0 and the $50B Game-Ad Market
All-In with Chamath Palihapitiya, Jason Calacanis, David Sacks and David Friedberg, and Adam Foroughi, co-founder and CEO of AppLovin
A roughly $50B mobile-gaming ad market with AppLovin’s own platform growing about 60% a year, the buyback program begun at the lows that has retired 20-25% of the shares for $6B, why discovery beats search against LLMs and agentic shopping, no precise-location tracking or social-graph data, and the 84% EBITDA margin he calls a moat.Ayar Labs CEO Mark Wade on the Interconnect Bottleneck, Co-Packaged Optics and the Road to an IPO
The Circuit with Ben Bajarin and Jay Goldberg, and Mark Wade, CEO, and Jesse Leiter, VP of Capital Strategy and Investor Relations, Ayar Labs
Why AI revenue growth is now capped by scale-up interconnect, the TSMC COUPE bet with a 3nm electronic IC, 2 and 4 Tb/s chiplets already in customer labs with 8 Tb/s coming, $650M raised this year plus a secondary, NVLink Fusion attach as the real TAM, and fiber-to-chip coupling as the yield crunch.SemiAnalysis: 300 Local Data-Center Bans, Three Projects Delayed
SemiAnalysis Weekly with Jordan Nanos, and Maya Barkin and Reyk Knühtsen of SemiAnalysis’s data-center, energy and industrials team
They mapped 400-plus moratorium instruments across 17 states against a 3,500-project pipeline: 77 campuses inside restricted jurisdictions, three actually delayed by local moratoriums, 80% of bans covering zero capacity, New York’s statewide freeze slowing roughly 800MW as the one that bites, and SpaceX’s Brownsville de-annexation.Positron’s Thomas Sohmers on Memory-Bound Inference, Frontier-Lab Margins and Whether the Data Centers Get Built
20VC with Harry Stebbings, and Thomas Sohmers, co-founder and chairman of Positron AI
GPU FLOPs up 120x since 2014 against 17x for memory bandwidth, cached tokens at roughly a thousandth of fresh compute, by his estimate, behind the labs’ margins, the $60-to-under-$1 collapse in token prices against what he estimates is a 100-1000x rise in intelligence per token, his claim that the anti-data-center backlash is a Chinese psyop, and why he fears concentrated power more than fast AI.Robinhood CEO Vlad Tenev on Tokenizing Everything, Trump Accounts and AI Agents That Won’t Trade
Moonshots with Peter Diamandis, Dave Blundin and Alexander Wissner-Gross, and Vlad Tenev, co-founder and CEO of Robinhood
By Tenev’s account: Robinhood Chain’s 200 stock tokens tracking US stocks and $1B-plus daily DEX volume, sole initial broker for the Trump Accounts program (7M opened, $1,000 per child), 100,000-plus accounts running AI trading agents that often refuse to trade, and retail access to OpenAI and Crusoe through the Robinhood Ventures funds.Inside Gemini: How Google Runs Product for Its Model
The Skip with Nikhyl Singhal, and Tulsee Doshi, product lead for Google’s Gemini models
Gemini 4 in development with the internal hype building, evals over PRDs and the 20-prompt mini-eval, why stack-ranking requests yields a boring product, live-experiment behavior shifts as the ship bar, Antigravity dogfooded by PMs, and why the app leads the API.
Read the Brief on Ethmos | Listen | Watch
