Why Publishers Are Licensing Content To AI Companies

For years, publishers fought for traffic.

Now many are fighting for compensation instead.

Because AI search is changing the economics of the internet itself.

ChatGPT, Google AI Overviews, Perplexity and other AI systems increasingly answer questions without sending users to the original source. That creates a major problem for publishers: fewer clicks, fewer ad impressions, weaker referral traffic and declining search visibility across the board.

At the same time, AI companies desperately need something publishers already have: high-quality content.

That is why a new market is emerging at extraordinary speed — AI content licensing.

And in 2026, some of the world’s largest publishers are now signing multimillion-dollar licensing deals with AI companies, even while simultaneously suing others.

The old internet deal is breaking

For decades, the relationship between publishers and search engines was relatively clear.

Google indexed content, publishers received traffic and traffic generated advertising revenue. That exchange powered much of the modern web.

AI changes the model completely.

Instead of directing users to articles, AI systems increasingly summarise information, generate answers, synthesise reporting, quote source material and provide recommendations directly inside the interface itself.

Users increasingly get what they need without clicking through.

According to Bain & Company, around 60% of searches now end without users progressing to another website. (Presenc AI)

For publishers, that changes everything because AI search weakens the traffic model the web depended on for years.

Publishers are already seeing traffic decline

The pressure is already visible across the media industry.

Reuters Institute research covered by The Guardian found publishers expect search traffic to decline by 43% over the next three years because of AI summaries and chatbot discovery. (BuzzStream)

Google referrals to publishers are reportedly already down significantly in some sectors.

This is why publishers increasingly view AI companies in two ways simultaneously: as existential threats and as potential revenue partners.

That contradiction now sits at the centre of the licensing boom.

AI companies need trusted content

AI models are only as useful as the information they can access.

And publishers produce something AI systems value enormously: professionally edited reporting, trusted journalism, structured information, real-time news coverage, domain expertise and historical archives built over decades.

That content improves model quality, factual grounding, answer freshness, citation reliability and overall user trust.

As Presenc AI noted in its 2026 licensing research, the bilateral licensing market has evolved into a recognisable commercial layer involving training rights, live data feeds, attribution agreements, real-time content access and per-use licensing models. (Presenc AI)

This is no longer experimental.

It is becoming infrastructure for AI systems.

The licensing deals are getting bigger

The scale of recent agreements shows how valuable publisher content has become.

News Corp and OpenAI

In 2024, News Corp signed a five-year agreement with OpenAI reportedly worth more than $250 million. (New York Post)

The deal covered content from The Wall Street Journal, The Times, The Sunday Times, New York Post and Dow Jones properties.

The agreement included training access, real-time content usage and AI product integrations.

Then in 2026, News Corp signed another major licensing agreement with Meta reportedly worth up to $50 million per year. (MEDIANAMA)

News Corp CEO Robert Thomson described the company as:

“An AI input company.” (The Guardian)

That statement matters because it reframes publishers as infrastructure providers for AI systems rather than just media businesses.

Axel Springer and OpenAI

Axel Springer, publisher of Politico and Business Insider, signed a multiyear OpenAI deal reportedly worth “tens of millions of euros”. (AI Watch.dog)

The agreement allowed OpenAI to train on publisher content, integrate summaries into ChatGPT and provide attribution links back to original reporting.

This became one of the earliest major publisher-AI partnerships, and many others followed soon after.

Condé Nast, Vox and The Atlantic

OpenAI also signed agreements with Condé Nast, Vox Media, The Atlantic, Associated Press and Le Monde, among others. (Presenc AI)

The Associated Press deal alone reportedly gave OpenAI access to archived news content dating back to 1985. (AI Watch.dog)

The pattern is becoming increasingly clear: AI companies want premium, trusted and professionally produced information at scale.

Publishers are trying to regain leverage

One major reason licensing deals are accelerating is because publishers are trying to establish economic leverage before AI systems become fully dominant.

Many executives fear that once AI companies possess enough training data, publishers may lose negotiating power permanently.

That is why publishers increasingly pursue a dual strategy: license selectively while litigating aggressively.

News Corp is a perfect example.

The company has signed major licensing agreements with OpenAI and Meta while simultaneously pursuing legal action against other AI companies over alleged unauthorised scraping. (AdExchanger)

The New York Times has also sued OpenAI and Microsoft over copyright claims while broader licensing negotiations continue across the industry. (A Media Operator)

The message from publishers is becoming increasingly direct:

“If AI companies want our content, they need to pay for it.”

AI licensing is becoming its own market

In 2026, the industry is rapidly moving beyond one-off deals.

Entire marketplaces are now emerging around AI content licensing.

Microsoft recently announced development of a Publisher Content Marketplace designed to help publishers license content, define usage rights, control crawler access and monetise AI training usage. (The Verge)

The platform is being developed alongside partners including Vox Media, Associated Press, Condé Nast and Yahoo.

Meanwhile, Dow Jones’ Factiva business reportedly now offers AI rights access to more than 8,100 content sources. (Wall Street Journal)

This signals something important.

AI licensing is evolving from isolated agreements into a scalable economic layer for the web itself.

Smaller publishers face a tougher challenge

Large publishers can negotiate major deals. Smaller publishers often cannot.

That is creating a growing divide inside the publishing industry.

Major media companies possess legal leverage, premium archives, recognised brands and negotiation power. Smaller publishers risk being scraped without compensation, excluded from licensing agreements and displaced by AI-generated summaries.

This is one reason new standards and protections are beginning to emerge.

Publishers are trying to control AI crawlers

One of the most important developments in 2026 is the rise of AI crawler licensing systems.

The Really Simple Licensing (RSL) standard allows publishers to embed licensing instructions directly into robots.txt infrastructure. (Wikipedia)

Participating companies include Reddit, Yahoo and Medium.

The idea is simple: publishers define usage rights, AI crawlers read licensing rules and AI companies either comply or negotiate access.

This represents a major shift in how the web functions.

The internet is slowly moving from:

“crawl first, ask later”

towards:

“license first, crawl later.”

Reddit shows how valuable community content has become

Reddit’s strategy highlights another important trend.

AI companies increasingly value forums, discussions, user-generated content and conversational language because this data improves how AI systems communicate naturally.

Reddit signed major licensing agreements with Google and OpenAI before later suing Anthropic over alleged unauthorised scraping. (AP News)

Again, the pattern repeats:

monetise authorised access
fight unauthorised access

This is rapidly becoming the default publisher strategy in the AI era.

Publishers are becoming infrastructure for AI

This may be the biggest shift of all.

Publishers are no longer just creating content for human readers. Increasingly, they are producing training data for AI systems.

That changes how publishers think about archives, syndication, licensing, attribution, monetisation and digital discovery itself.

In many ways, publishers are becoming suppliers inside the broader AI economy.

Robert Thomson’s phrase:

“AI input company”

may ultimately define the entire industry transition. (The Guardian)

The future of publishing may depend on licensing

The uncomfortable reality is this: AI search is weakening traditional traffic economics, and publishers know it.

That is why licensing deals are accelerating so quickly.

Many publishers increasingly believe the future may depend more on licensing revenue, AI partnerships, controlled access, attribution systems and crawler monetisation than search traffic alone.

The web’s old economic model was built on clicks.

The next one may be built on data access.

The AI economy is rebuilding the internet’s value chain

This is bigger than publishing.

It represents a complete restructuring of how value flows across the internet.

AI companies need trusted information, real-time updates, authoritative sources and conversational training data. Publishers need compensation, leverage, sustainability and visibility.

Licensing deals are the compromise currently emerging between those two forces.

But the long-term outcome is still unclear.

What is clear is this:

The relationship between AI and publishing is no longer theoretical.

It is becoming one of the defining economic negotiations of the AI era.