So, here’s a cool approach we’re seeing from media and info companies. They are now turning their intellectual property (IP) into a revenue stream by licensing it to those creating large language models (LLMs) for AI tools.
This was one of the key insights shared by Laura Martin, Senior Entertainment & Internet Analyst at Needham & Company, during her keynote address at the Parks Associates Future of Video conference held in Southern California.
For instance, CuriosityStream is a standout example of a company cashing in on this trend. They started licensing their collection of factual documentaries and TV series to firms involved in developing AI technologies. This content is essential for training LLMs, providing them with reliable information that can be spun into services, such as chatbots.
Thanks to its clever decision to license content to AI developers, CuriosityStream saw a massive 46% rise in its overall revenue, hitting $18.4 million, while also reducing losses to below $4 million in the third quarter of the year, as previously noted by The Desk.
Interestingly, licensing isn’t the only income stream for CuriosityStream; their main gig remains distributing content via their own premium streaming service, alongside free, ad-supported channels around the U.S. and elsewhere.
However, Martin highlighted that aiding in the growth of AI language models, rather than shying away from it, played a pivotal role in the surge of their revenue during Q3. She strongly recommends that other companies follow CuriosityStream’s footsteps.
According to Martin, “In our view on Wall Street, any entity managing first-party data should work on generating income from licensing it for large language models. If you’re not in the first-party data game, you’re likely to fall behind, as those reaping rewards from LLMs will outpace you.”
For weeks, Martin has consistently pointed to this notion. Her comments echo those made in a recent debate hosted by Streaming Media, where she again pushed the idea that selling data to AI developers ranks among two new vital revenue streams for media companies. The second, which has captured the spotlight right now, is shoppable TV.
She has quipped, “Multiplying your revenue streams can significantly enhance your market valuation. Adding new streams could boost your company’s valuation by 50% per licensed stream.”
Many large media companies aren’t quite onboard with proactive licensing of their first-party data and other content, especially relating to characters or shows. For instance, the Walt Disney Company and Universal Pictures recently went for a lawsuit against Midjourney, as some users were using the platform to create images eerily similar to beloved characters like Winnie the Pooh and Shrek. Warner Bros Discovery followed suit with a related claim three months later, accusing Midjourney of exploiting their content for developing their LLM without any permission or financial compensation.
But it’s not just entertainment firms grappling with AI developers. Last year, the New York Times took legal action against OpenAI, the company behind ChatGPT, stating that OpenAI violated copyright by incorporating their articles into its generative AI knowledge base. Both the AI developer and a key investor, Microsoft, disputed these claims.
Still, it seems these companies are open to negotiating licensing agreements, given they are compensated appropriately and there are sufficient safeguards in place. Some firms, like Reddit, have already brokered deals with AI developers to license their assets.
Now CuriosityStream joins those charging ahead with licensing solutions for LLM developers. CEO Chris Stinchcomb points out that they can comfortably dive into these arrangements typically associated with news agencies and publishers, given their offerings consist of authentic documentaries and factual series instead of fictional content.
Moreover, they face fewer issues with actor payments and unions, as their content is more fact-oriented. This makes it a breeze to strike licensing deals, particularly since their materials boast extensive and detailed metadata.
Stinchcomb shared, “Navigating the nonfiction category is definitely simpler than dealing with scripted media, which brings along a whole set of talent and union considerations.”
This post is part of our in-depth coverage surrounding the Parks Associates Future of Video conference, which unveiled the “State of Streaming 2026” report, available for free download here.To view more posts related to this event, click or tap here..
