Every major AI lab publishers negotiate with — OpenAI, Google, Meta, Microsoft, Amazon, Perplexity — has signed content licensing deals. Anthropic has not, and neither has xAI. What makes Anthropic the more interesting case is what hasn’t happened as a result: no digital publisher has sued it. Compare that to the rest of the field — OpenAI faces suits from The New York Times and Ziff Davis, Perplexity from News Corp. and CNN, Cohere from Condé Nast, and Google from Penske Media. Anthropic did settle a $1.5 billion class action from book authors this year. That’s a different plaintiff class entirely, and publishers haven’t followed the same path.
Anthropic’s own explanation, per a company statement, is that Claude trains on “a combination of publicly available data, data we create and data obtained through specific partnerships, based on product needs” — a position built on the company’s belief that content scraping is legally permissible under fair use. It has simply not pursued licensing deals on that basis. Tom Turvey, Anthropic’s VP of product partnerships since 2024, was described by one publisher executive as being there to “pick up the phone twice a day and say no.” The company reportedly turns down content offers unless the material is genuinely hard to access — one cited example was a half-million Dutch medical journals.
01Why this matters for publishers and buyers
- Anthropic isn’t playing the same negotiation game as its peers. OpenAI, Google, and Meta signing deals created a market rate publishers could point to; Anthropic opting out of that market denies publishers a comparable, and that changes the leverage calculus in every other negotiation.
- The lawsuit gap is not proof of a stronger legal position — it may just be lower exposure. Anthropic’s enterprise-first strategy means less of its training and product surface touches consumer-facing publisher content directly, which is a narrower target than a mass-market answer engine scraping news for search-style summaries.
- The fair-use bet is the actual story for every publisher weighing whether to block AI crawlers. If Anthropic’s position holds without a publisher suit testing it, that’s a signal other AI firms will read as license to make the same bet — the opposite of the licensing-market Anthropic’s competitors are building.
- Anthropic’s enterprise focus is already commercially significant. It’s reported to be capturing 73% of first-time AI tool spending among companies, per March data cited by Axios. That scale doesn’t depend on publisher content deals to keep growing.
02What publishers should do
03The bottom line
Anthropic’s zero-deals, zero-suits record isn’t a sign that AI licensing pressure is easing — it’s a sign that one major lab is betting it can sit outside that market entirely, and so far nobody has forced the question in court. Every operator’s AI-crawler and licensing strategy should account for that gap explicitly, rather than treating “the AI companies” as one negotiating counterparty with one set of incentives.