CoreWeave Crash, AI Trade Cracks, Anthropic IPO Red Flags, Bill Gates Doomer
Oct 17, 2026 · 1h 34m
More Technology podcasts · More Business podcasts
Summary
Ben Horowitz joins to discuss the neocloud debt crisis, analyzing CoreWeave’s financial structure and rising bond yields amid a GPU shortage. The panel debates whether this signals an AI bubble or normal volatility, noting that demand remains high despite deflationary software trends. They also explore the critical importance of power infrastructure and payback periods as key metrics for sustainable AI growth.
Topics discussed
Welcome and introductions of guests David Sachs and Ben Horowitz
CoreWeave debt yield spike and neocloud stock decline
Explaining CoreWeave's box economics and independent financing
Credit derivative swaps and rising default risk metrics
Rising interest rates and the $5 billion breakeven risk
Ben Horowitz on GPU duration, supply shortage, and tailwinds
Comparison to past bubbles and the danger of arbitrage businesses
Is the CoreWeave crash volatility or the start of a bubble pop?
Evidence of massive compute demand and supply constraints
Project-level credit backing and secondary market bond pricing
Customer concentration risk and the venture industrial complex
OpenAI and Anthropic IPOs and defensive portfolio positioning
The deflationary nature of AI and software margins
The future of software development and coder demand
Sustainability of software franchises and new moats
Power as a moat and the importance of payback periods
Risks of extended payback periods and credit default management
NVIDIA's control over HBM and the case for long energy
Anthropic's revenue, IPO potential, and Andreessen's position
OpenAI's transition from nonprofit and investor day reactions
PSYOP Corner: Anthropic's valuation strategy and war chest
Ben Horowitz on cash reserves and compute-backed debt
Foundation model duopoly and advice for Elon Musk
Promotion for the election night live stream
Anthropic's claims about AI-powered IoT worms and panic
Distillation attacks and the failure of government intervention
Technical reality of AI worms and IoT security flaws
Anthropic's security negligence and the irony of their warnings
The trend toward small, customized enterprise models
Critique of Anthropic's black hat research and PR strategy
AI consciousness, moral patients, and the hall of mirrors
Bill Gates' role in AI doomerism and gain-of-function research
Bill Gates' political influence and motivation for fear-mongering
The four requirements for continuous AI progress and why they fail
Final thoughts on building vs. doom and closing remarks
More from All-In with Chamath, Jason, Sacks & Friedberg
Anthropic Goes Public, Nvidia in Crisis, Markets Tank, Open Source Crackdown?
Nov 20, 2026
Databricks CEO: AGI Is Already Here, Why Companies Are Behind, and the Superintelligence Debate
Nov 17, 2026
Jensen Ousted as Nvidia CEO: "Spy" Scandal, Export Restrictions, Market Impact
Nov 14, 2026
Freedom vs. Safety: Axon CEO Rick Smith on the Future of Policing
Nov 10, 2026
Midterm Blood Bath, $4T AI Sell Off, Data Center Freeze, What's Next for Trump
Nov 7, 2026
Listen ad-free on Castria