Which Media Companies Profit When AI Pays for Archives

Two regional newspapers filed a lawsuit in federal court on Friday that sounds, on its surface, like another round in a familiar fight. The Seattle Times and Newsday sued OpenAI and Microsoft, alleging the tech companies copied their journalism without permission to train their AI systems. The complaint, filed in the U.S. District Court for the Southern District of New York, alleges that OpenAI and Microsoft scraped the newspapers’ websites, including content behind paywalls, and incorporated articles into datasets used to train and operate products including ChatGPT, Microsoft Copilot, and Bing’s AI features. Microsoft’s response was characteristically understated: a spokesperson said the company was “surprised by the lawsuit” and was “always happy to sit down and explore solutions to this type of dispute.”

What makes this filing more consequential than the ones before it is the remedy sought. The publishers are asking for an order demanding the destruction of copies of their copyrighted works, as well as the training datasets and AI models incorporating that material. Destruction, not just damages. If a court were ever to grant that, the cost to OpenAI and Microsoft would dwarf any licensing check. And it is that asymmetry, not the lawsuit itself, that should focus the attention of long-term investors.

The question worth asking is not whether the Seattle Times wins. It is what happens to the value of journalistic archives if courts gradually force AI companies into publisher-by-publisher licensing arrangements. The answer is already visible in deal flow.

OpenAI has signed publisher and data deals, the single largest being a reported deal that could be worth more than $250 million over five years with News Corp, based on an estimate reported by The Wall Street Journal when the partnership was announced in May 2024. Meta Platforms has separately signed a multiyear AI content licensing deal with News Corp that could be worth up to $50 million a year, also reported by The Wall Street Journal. That agreement covers content from News Corp’s U.S. and U.K. media properties, including titles such as The Wall Street Journal and the New York Post, as well as U.K. papers including The Times and The Sunday Times. News Corp (NWSA) did not stumble into this position. It negotiated from a portfolio of well-documented, rights-cleared archives that smaller regional publishers simply cannot replicate at scale.

The New York Times (NYT) is playing a different hand. Earlier this week, the U.S. Justice Department filed a brief in The New York Times’ case against OpenAI and Microsoft that backed the companies’ fair-use arguments and said the benefits of training AI models on such content outweigh potential harms. That intervention helps OpenAI and Microsoft in the short run but does not resolve the underlying pressure on the industry. The Times has already spent more than $28 million on litigation costs against AI companies, with $4.2 million coming in the first quarter of 2026 alone, according to the company’s filings. Meanwhile, the Times struck a multiyear content-licensing deal with Amazon for AI-related uses and is developing its own internal use cases for the technology. It is fighting and licensing at the same time, which is probably the right strategic posture.

The harder look should go to the publishers complaining loudest without a comparable asset base. The Seattle Times and Newsday complaint says referral traffic from search has fallen sharply for publishers, and cites industry data to illustrate the scale of the decline. That is a genuine wound. But filing suit does not automatically create licensing leverage. Many industry analyses have argued that the long tail of small and mid-size publishers is likely to see limited revenue from AI licensing. The structural advantage belongs to publishers with scale, cleared rights, and continuous output that AI companies want as an ongoing feed, not a one-time archive dump.

The market is already shifting from one-time training purchases to live-access deals, with some industry trackers counting attribution and live-access agreements rising from two in 2023 to a projection of about 34 in 2026. AI companies increasingly want ongoing feeds and real-time data, not static archives.

News Corp holds that. The New York Times holds it. Most of the publishers joining this litigation wave do not. The Seattle Times and Newsday may win sympathy, and perhaps a settlement check. But the compounding asset, the one that keeps generating licensing income as model after model gets built, belongs to the scaled owners of content that AI cannot do without. That distinction is the one worth tracking.