Anthropic CEO warns of AI “tsunami” as power and wealth concentrate in a few hands

Summary

Anthropic CEO Dario Amodei has compared the rapid advance of artificial intelligence to a “tsunami” on the horizon and warned that society is dangerously underestimating how close human-level systems may be. In recent interviews, including a long-form podcast with Indian investor Nikhil Kamath, he said AI’s economic, geopolitical and safety consequences will be “enormous” and argued that the technology needs to be actively steered in the right direction rather than left to a handful of corporate leaders.

In practice

Amodei frames recent progress as a story of scaling laws: as companies pour in more data, computing power and model parameters, AI systems steadily gain more general problem-solving abilities. For workers and organisations, this points to tools that can take over increasingly complex tasks, from writing and coding to analysis and decision support, potentially reshaping how businesses operate and how countries compete.

On the business side, Anthropic itself is now one of the most prominent players in the sector, with fast revenue growth and a multi-billion valuation. At a recent event in India, Amodei described a near future where AI agents operate like “a country of geniuses in a data centre”, coordinating at superhuman speed, and suggested that coding may be among the first skills to be widely automated, leaving critical thinking as humanity’s “last real edge”.

Who benefits / who loses

Amodei is candid about the new concentration of power and wealth around a small cluster of AI firms, saying he is “somewhat uncomfortable” with how quickly it has happened “almost overnight, almost by accident”. He has warned that AI could create trillionaires and trigger backlash if most of the gains flow to a tiny elite, arguing that you cannot simply promise abundance while keeping a disproportionate share of the rewards for those building the systems.

At the same time, he calls for governments and the public to have a say in how AI develops, even though stronger rules may go against short-term commercial incentives for companies like Anthropic. He worries that technical work on safety and alignment is moving faster than public awareness or government action, creating a gap where highly capable systems arrive before societies have decided how they should be controlled, shared and integrated into everyday life.

Why it matters

• The people building frontier AI are now among the loudest voices warning about its risks and demanding guardrails.
• Rapid automation of skills like coding could shift which jobs are safe and which careers need rethinking in the near term.
• Extreme concentration of AI power and profits may deepen inequality and fuel political backlash if not addressed.
• Governments’ slow response increases the chance that rules, protections and safety norms arrive only after major disruptions.