Descriptions:
The All-In Podcast brings together Chamath Palihapitiya, David Sacks, and investor Brad Gerstner to work through four interconnected stories that defined a consequential week in AI. The episode’s anchor is the Anthropic-SpaceX deal: Anthropic has agreed to take over the full capacity of Colossus 1, the data center built by XAI, which is simultaneously being folded entirely into SpaceX. The panel reads this as Elon Musk acknowledging that his competitive advantage in AI lies in infrastructure rather than model development — a significant strategic repositioning with long-term implications for who controls AI compute at scale.
The conversation expands into monopoly risk, with debate over whether Anthropic’s rapid ascent and enterprise dominance is creating an unhealthy concentration of market power, and whether regulatory action would help competitors or simply entrench incumbents through compliance costs. A substantial segment covers the Trump administration’s reported consideration of an ‘FDA for AI’ — an executive order to create a review process for new models, reportedly triggered by internal alarm over Anthropic’s Mythos model. Sacks and Gerstner largely frame this as political cover rather than coherent policy, though the debate surfaces genuine tensions around frontier model safety and government oversight.
The episode also revisits Terraab, Musk’s Texas chip manufacturing project, where new legal filings now put the cost at $55–119 billion — far exceeding earlier estimates of $20–25 billion. For investors and operators tracking both the competitive dynamics and policy environment surrounding AI infrastructure, this episode offers a pointed and well-informed set of perspectives.
📺 Source: All-In Podcast · Published May 08, 2026
🏷️ Format: Podcast
