Oct 28, 2026 · 45 min
Anthropic’s losses test AI’s business case
Anthropic's IPO Filing Reveals $80 Billion Loss. It Doesn't Matter
The episode connects Anthropic’s IPO ambitions to a broader question: whether AI investment, adoption, and modernization can produce durable gains beyond technology companies.
- 1Anthropic’s massive losses matter less than whether Claude can create durable value within an interdependent AI economy.
- 2Apollo’s productivity skepticism exposes the gap between individual time savings and measurable gains for entire organizations.
- 3COBOL modernization requires recovering undocumented institutional knowledge, not merely translating legacy code into a newer language.
Don't miss
The COBOL segment reveals why translating legacy code is easier than recovering the undocumented institutional knowledge that makes it work.
The brief
Anthropic’s IPO filing puts huge losses, ambitious projections, and extensive risk disclosures under scrutiny—but the deeper question is whether Claude can generate durable business value.
The hosts trace Anthropic’s ties to major companies and the AI sector’s stated obligations, asking whether interconnected valuations reflect real growth or a bank-run-like vulnerability.
Apollo’s skepticism shifts the focus from impressive individual use cases to organizational outcomes: saved minutes do not automatically become higher revenue per employee or broader productivity.
The COBOL discussion supplies the sharpest warning: AI may refactor old code, but decades of undocumented operational knowledge can make narrow automation dangerous.
Across valuation, productivity, and legacy systems, the episode argues that AI’s hardest test is converting technical capability into dependable institutional value.
Weekly email
State of the Pod, every Monday
The week's most-discussed names and stories from podcasts, with the moments behind them. Free, five minutes, unsubscribe in one click.