Open-Source AI Battle, Google Throttles Meta, Micron Margins Moon, Rocket Lab buys Iridium

Watch on YouTube ↗  |  June 29, 2026 at 20:27  |  2:25:39  |  TBPN
Speakers
John Coogan — Co-Host, TBPN
Chad Rigetti — Founder, Rigetti Computing
Pim — Co-founder & CEO, General Intuition
Jordi Hays — Co-Host, TBPN
Yadin Sofer — Co-founder & CEO, Tracer
Chris Altek — Founder, Cadence
Neil Mova — Co-founder & CEO, Sale Research

Summary

The episode reviews the renewed open-source AI debate after China's GLM 5.2, Google limiting Meta's Gemini access, and Micron's surging memory profits. It also covers Comcast's split, European World Cup tourism, and startup interviews across privacy AI, quantum computing, robotics, inference, defense, healthcare, and venture capital.

  • Chinese open-weight model GLM 5.2 reopens the open-source vs closed-source AI security and commoditization debate.
  • Google limits Meta's Gemini capacity, signaling scarce AI compute and bullish Google Cloud demand.
  • Micron, Samsung, and SK hynix are seen as huge beneficiaries of AI-driven memory price surges.
  • Comcast announces a planned split of connectivity and media businesses.
  • Interviews cover quantum AI servers, local privacy models, clinical AI, inference optimization, robotics, and subterranean defense.
  • Venture capital discussions focus on LA hard-tech, El Segundo talent, and shifting startup categories.
Ideas
John Coogan Co-Host, TBPN 26:14
Google capacity scarcity is bullish
Google limiting Meta's access to Gemini capacity shows AI compute is a scarce commodity and demand is outstripping Google's capacity. This is extremely bullish for Google because it confirms Google Cloud acceleration and pricing power in AI services, even though Google does not break out Gemini revenue.
John Coogan Co-Host, TBPN 39:14
Memory makers profit from AI shortage
Micron's explosive profit growth shows an enormous transfer of cash from AI providers to memory chip makers. Micron raised DRAM prices more than 60% sequentially and NAND prices more than 80%, while data center demand and limited high-bandwidth memory capacity keep supply tight. Micron, Samsung Electronics, and SK hynix are analogized to oil producers for the AI airline industry; Micron joined the trillion-dollar club and the Barron's bull case says the stock could still double.
Chad Rigetti Founder, Rigetti Computing 67:52
Quantum co-processors may accelerate AI training
Quantum computing is moving from science project to AI data center co-processor. Multimodal quantum hardware can target several orders of magnitude speedup for key AI training workloads, reducing cost and power. Chad expects the quantum mainstream inflection when quantum systems run production data center workloads without needing the 'quantum' label, potentially within five to seven years.
Pim Co-founder & CEO, General Intuition 93:18
Robotics supply chain may favor game inputs
Models trained on roughly a trillion action tokens from game-controller inputs can control any robot that already ships with game-controller interfaces. Because intelligence is the bottleneck, the robotics supply chain will converge on gaming inputs instead of humanoid robots in the next two years.
Up Next

This TBPN video, published June 29, 2026, features John Coogan, Chad Rigetti, Pim discussing GOOG, MU, 005930.KS, 000660.KS, QUBT, Humanoid robots. 4 trade ideas extracted by AI with direction and confidence scoring.

Speakers: John Coogan, Chad Rigetti, Pim  · Tickers: GOOG, MU, 005930.KS, 000660.KS, QUBT, Humanoid robots