Publishers want billions from AI platforms

Two decades ago, the news industry handed its content to search engines and social platforms for free. It took nearly twenty years of painful paywall building to recover from that mistake, and publishers are determined not to repeat it with AI.

Major media giants including IAC, The New York Times, News Corp, and Axel Springer are close to forming a joint coalition to push back against AI companies scraping their reporting without consent. According to a report from Semafor, the group is weighing major lawsuits and pushing lawmakers for strict legislative action.

Search found the web; LLMs are stripping it

IAC Chairman Barry Diller has been outspoken on this front, arguing that AI engines are simply cannibalizing publisher traffic by vacuuming up original reporting and spitting it back out as a repackaged answer.

IAC CEO Joey Levin, who is playing a pivotal role in organizing the coalition, put it bluntly: while everyone is obsessed with sci-fi fears of AI wiping out humanity, AI taking over media is far more immediate and dangerous.

Levin noted that while traditional search was built to guide users toward the best of the web, generative AI feels designed to simply strip the best of the web for itself.

An eight-figure offer vs. a multi-billion expectation

Tech companies know they need licensing deals, but the financial gap between the two sides is massive.

According to reports, tech firms are hoping to placate publishers with payouts somewhere in the tens of millions of dollars (eight figures). Publishers, however, say that if generative models rely heavily on professional reporting for their reasoning and accuracy, compensation across the industry needs to match that value—reaching into the billions of dollars.

Tech companies are trying different tactics to de-escalate the tension. Microsoft's consumer chief Yusuf Mehdi argued the company wants to drive traffic through Bing and Edge chat interfaces, promising future ad revenue sharing inside Bing Chat.

OpenAI took a divide-and-conquer approach, recently signing a licensing deal with the Associated Press for access to its text archives in exchange for technology access. Google, meanwhile, is pitching AI-assisted writing tools directly to newsrooms.

Why it matters

We think the media industry is playing this very smartly. By forming a collective front instead of negotiating fragmented, one-off deals, publishers avoid being picked apart one by one. They learned the hard way with search and social that once distribution and interface control are lost, publisher leverage vanishes completely.

The real battleground here is valuation. An eight-figure peace offering from Big Tech compared to a multi-billion-dollar publisher demand shows neither side is close. This almost guarantees prolonged legal battles and aggressive lobbying pushes.

For builders and the wider AI sector, this is the first serious crack in the free training data assumption. If publishers successfully force Big Tech into recurring, multi-billion-dollar licensing structures, the unit economics of training next-generation frontier models will change dramatically, pushing AI development costs higher and widening the moat between tech giants and resource-strapped open-source projects.