Program and Guests
A reposted interview by Live Chart Analysis, featuring Anthropic CEO Dario Amodei and his sister, co-founder Daniela Amodei. Original English title: Anthropic CEO BREAKS SILENCE! Exclusive Interview on the Future of AI. This is a repost, apparently of a CNBC Squawk Box interview on IPO day; the opening introduction is missing, but the main Q&A is complete.
Key Takeaways
- Daniela: The company will become the largest publicly traded public benefit corporation; the law requires a balance between shareholder interests and the public interest. 0:52
- Frontier competition is about quality, not price: each step up in cognitive capability brings much greater economic value, and value is concentrated at the frontier. 5:04
- Haiku 5.5 cut its price by a factor of 4 compared with Haiku 4.5; the same technology can be used both to build frontier models and to make smaller, cheaper models. 5:40
- The 900 billion-dollar commitment is spread over several years. Compared with the cash on the balance sheet and the revenue already generated each year, “suddenly it doesn’t seem so large.” 7:15
- Building data centers after the midterm elections may be restricted, but the company pledges to bear the grid price differential and not pass it on to consumers. 6:43
Background and Notes
A reposted interview by Live Chart Analysis in which Dario Amodei and Daniela Amodei discuss the 900 billion-dollar compute commitment, Haiku price cuts, frontier competition, and the public benefit corporation structure. This is a repost, apparently of a CNBC Squawk Box interview on IPO day; the opening introduction is missing, but the main Q&A is complete. The repost must not be described as an original publication. Subtitle coverage note: the subtitles for this video contain 452 timestamped segments, covering 0:00–14:53 (total length 14:53), with full-video coverage; this article compiles key sections of the interview from 0:52–7:15, and does not cover sections after 7:15. The views and timestamps cited are based on the original subtitles; each timestamp marks the start of the corresponding spoken segment. Publication date note: the publication date is the Hong Kong date 2026-11-20, with day-level precision.