Anthropic Balances Growth, Costs, and Brand Protection
Sep/30/2025
Anthropic, the AI startup behind the Claude models, has prevailed in a WIPO arbitration over the domain anthropic.icu, registered in 2024 and used to host pay-per-click ads referencing its products. The panel ruled that the site’s use of the Anthropic mark — registered in the United States in 2024 — constituted bad - faith registration and exploitation of the company’s goodwill, ordering the domain transferred to the firm. Founded in 2021 by former OpenAI staff, Anthropic has swiftly risen to the front ranks of generative AI with its Claude models, the latest of which — Claude 3 — comes in three versions, from the lightweight Haiku to the heavyweight Opus, already integrated into Amazon’s Bedrock cloud platform. Backed by nearly $4 billion from Amazon and up to $2 billion from Google, the startup positions itself as a safer, more principled AI developer, even embedding the Universal Declaration of Human Rights into Claude’s training framework. Yet, like its rivals OpenAI and Google, Anthropic faces the heavy economics of AI: costs can run to a billion dollars a year, while revenues remain modest.
