Meta Platforms has struck a multimillion-dollar licensing agreement with News Corp, the media empire controlled by Rupert Murdoch, in a deal that signals a significant shift in how technology companies compensate publishers for the use of their content in training artificial intelligence models. The arrangement, reportedly worth more than $100 million over multiple years, grants Meta access to articles and archival material from some of the world’s most recognizable news brands, including The Wall Street Journal, The New York Post, Barron’s, MarketWatch, and several major Australian and British publications.
The agreement comes at a time when the relationship between Big Tech and the news industry is being fundamentally renegotiated. Publishers have grown increasingly aggressive in demanding compensation for the vast troves of copyrighted journalism that AI companies have used—often without explicit permission—to train large language models. Meta, which operates the Llama family of open-source AI models, appears to be positioning itself ahead of potential legal exposure by securing formal rights to high-quality news content, as first reported by Engadget.
The Scope of the Deal and What Meta Gets
Under the terms of the agreement, Meta will gain licensing rights to content produced across News Corp’s portfolio, which spans continents and media formats. In the United States, that includes Dow Jones properties such as The Wall Street Journal, Barron’s, and MarketWatch, as well as the tabloid New York Post. In Australia, the deal covers mastheads under the news.com.au umbrella, The Australian, and The Daily Telegraph, among others. In the United Kingdom, News Corp’s holdings include The Times and The Sun, both of which are understood to be part of the arrangement.
The content will be used to train Meta’s AI products, which range from the company’s AI assistant integrated into WhatsApp, Instagram, and Facebook, to its broader Llama model development. For Meta, the deal provides a defensible legal foundation for incorporating journalistic material into its AI systems—a matter of growing urgency as copyright lawsuits from publishers mount across the industry. News Corp CEO Robert Thomson has been vocal about the need for fair compensation, and this deal appears to reflect that stance. Thomson previously described the unchecked scraping of news content by tech firms as a form of “kleptocracy,” a term he used in public remarks criticizing the industry’s practices.
News Corp’s Track Record of AI Licensing
News Corp is no stranger to these kinds of arrangements. The company signed a landmark deal with OpenAI in 2023, reportedly worth $250 million over five years, granting the ChatGPT maker access to current and archived articles from its publications. That agreement set an early benchmark for the value of premium news content in the AI training market and put pressure on other publishers to seek similar terms. The Meta deal, while reportedly smaller in total value, extends News Corp’s strategy of monetizing its journalism across multiple AI platforms rather than granting exclusivity to any single company.
The dual-deal approach gives News Corp a diversified revenue stream from AI licensing while also ensuring its content is represented in the outputs of competing AI systems. This is a calculated bet: by licensing widely, News Corp increases the likelihood that its journalism influences the factual grounding of AI-generated responses, which could in turn drive brand recognition and, potentially, subscriptions. It also hedges against the risk that any single AI platform will dominate the market.
The Broader Industry Context: Publishers Push Back
The Meta-News Corp deal arrives against a backdrop of escalating tension between publishers and AI developers. The New York Times filed a high-profile lawsuit against OpenAI and Microsoft in late 2023, alleging that millions of its articles were used without authorization to train AI models. That case remains ongoing and has become a bellwether for how courts may treat the application of fair use doctrine to AI training data. Other publishers, including The Intercept, Raw Story, and AlterNet, have filed their own suits against OpenAI and Meta.
Meta itself has faced direct legal challenges. In 2024, a group of authors and content creators sued the company, alleging that its Llama models were trained on copyrighted works obtained through pirated book databases. Meta has denied wrongdoing but has clearly taken steps to reduce its legal risk by pursuing formal licensing agreements. The News Corp deal is the most prominent example of this strategy, but it is unlikely to be the last. Reports indicate that Meta has been in discussions with multiple publishers and content providers about similar arrangements.
What This Means for Smaller Publishers
While deals of this magnitude are encouraging for large media conglomerates with the bargaining power to command nine-figure sums, the picture is far less clear for smaller and independent publishers. Many newsrooms lack the legal resources or negotiating leverage to secure comparable agreements, raising concerns about a two-tiered system in which well-funded legacy media organizations profit from AI while smaller outlets are left without compensation—or worse, see their content used without consent or attribution.
Industry groups such as the News Media Alliance have called for legislative solutions, including updates to copyright law that would create clearer frameworks for AI training data licensing. In the European Union, the AI Act and related regulations have begun to address some of these issues, requiring greater transparency about training data. In the United States, however, legislative action has been slower, leaving the market to be shaped largely by private negotiations and court rulings.
Meta’s AI Ambitions and the Role of Quality Data
For Meta, the investment in news licensing reflects a broader strategic imperative. The company has committed billions of dollars to AI development, with CEO Mark Zuckerberg describing artificial intelligence as the company’s single most important long-term investment. Meta’s Llama models, which are released as open-source software, compete with proprietary systems from OpenAI, Google, and Anthropic. The quality and breadth of training data is widely regarded as a key differentiator among these models, and access to professionally produced journalism offers a significant advantage in terms of factual accuracy and linguistic quality.
Meta’s AI assistant, which has been rolled out across its family of apps, is designed to answer user questions, summarize information, and generate content. Grounding these capabilities in verified, professionally reported news content could help Meta reduce the hallucination problem—the tendency of AI models to generate plausible-sounding but factually incorrect information. This is particularly important as AI-generated answers increasingly substitute for traditional web searches, a trend that threatens to further erode traffic to publisher websites.
The Economics of AI Licensing Are Still Taking Shape
The financial terms of these deals remain largely opaque, making it difficult to assess whether publishers are being fairly compensated relative to the value their content generates for AI companies. The reported $100 million-plus price tag for the Meta-News Corp deal is substantial, but it covers years of access to content from dozens of publications across multiple countries. Spread across the full portfolio and the duration of the agreement, the per-article economics may be modest.
Still, for an industry that has seen advertising revenues decimated by the rise of digital platforms—many of them operated by Meta itself—any new revenue stream is significant. News Corp reported total revenues of approximately $10 billion in its most recent fiscal year, so even a deal worth $100 million over several years represents a meaningful but not transformative contribution. The real question is whether AI licensing can scale into a durable category of revenue for the news industry as a whole, or whether it will remain confined to a handful of large players with the leverage to extract meaningful payments.
What Comes Next for Tech and Media
The Meta-News Corp agreement is likely to accelerate negotiations between other AI developers and publishers. Google, which has its own Gemini AI models, has already struck content deals with several publishers through its Google News Initiative and related programs. Anthropic, the maker of Claude, has been more circumspect but faces similar pressures. Apple, which is building AI features into its devices, has also reportedly engaged in licensing discussions with major news organizations.
For News Corp, the deal reinforces its position as the most commercially aggressive major publisher in the AI licensing market. Robert Thomson has consistently argued that the news industry must be compensated for the foundational role its content plays in training AI systems, and the company’s willingness to strike deals with multiple competing platforms suggests confidence that the market will support parallel arrangements. Whether other publishers can replicate this model—or whether the window for favorable terms will narrow as AI companies develop alternative data strategies—remains one of the most consequential open questions in the media business today.
WebProNews is an iEntry Publication