Can Anthropic afford to put safety first?
Nov 21, 2026 · 37m
More Technology podcasts · More Business podcasts
Summary
Nilay Patel and Hayden Field analyze Anthropic’s historic IPO, examining the tension between its public market obligations and its commitment to AI safety. They discuss the company’s complex governance structure, where seven founders hold unchecked control via a single share, and how this insulates them from shareholder pressure. The conversation also covers the massive compute costs Anthropic faces and the strategic implications of OpenAI allowing Anthropic to go public first.
Topics discussed
Sponsor: CrowdStrike and AI cybersecurity
Anthropic IPO overview and episode preview
Introduction of guest Hayden Field
IPO day performance and new millionaires
OpenAI's reaction to Anthropic going public
Business fundamentals: token costs and revenue targets
Corporate structure and shareholder power
Unit economics: renting capacity and margins
Customer backlash and the strategy of 'moats'
Sponsor: CrowdStrike AI Detection and Response
The dilemma of enterprise data access
SMBs creating internal tools for model routing
Earnings shakeout and supplier conflicts
Codifying safety commitments in governance
The Long-Term Benefit Trust structure
Board composition and founder control
The seven-founder voting mechanism
Sponsor: Agentic AI and security vulnerabilities
Safety incentives and the 'greater good' narrative
Founder beliefs and the danger of unchecked power
Comparison with NVIDIA's governance scandal
Silicon Valley founder insulation culture
Trust powers and board removal mechanics
Financial pressures and bankruptcy risk
Compute constraints and R&D investment
NVIDIA's impact on Anthropic's chip supply
What to watch in Anthropic's stock performance
Earnings calls vs. day-to-day hype
Closing remarks and show promotion
Sponsor: CrowdStrike enterprise security
More from Decoder with Nilay Patel
Listen ad-free on Castria