Since 2023, AI companies have signed 91 confirmed content licensing deals with news publishers, and the money involved is real: one deal alone is reported at up to 50 million US dollars per year. For brands and PR teams, that raises a practical question: if you pitch a publisher that has signed an AI licensing deal, does your coverage actually get cited more in AI search?
Together with Press Ranger, OtterlyAI cross-referenced 20M+ URLs cited across 7 AI Search platforms in June 2026 against every confirmed AI content licensing deal on record. This is the first study we know of that joins the deal side and the citation side at scale.
Key Findings(TL;DR)
News publishers with an OpenAI deal earn 48% more AI citations per page on ChatGPT than publishers with no deal (10.2 versus 6.9 citations per cited page). Across all 7 platforms the premium is 46%.
Only two platforms reward licensed publishers at all. ChatGPT (+48%) and Microsoft Copilot (+85%). On Perplexity, Google AI Overviews, Google AI Mode and Gemini, licensed publishers land at or below publishers with no deal.
The Copilot number is not a Microsoft effect. Microsoft has zero exclusive publishers, and 99% of Copilot’s citations to Microsoft licensees land on publishers that also signed with OpenAI. Strip that overlap out and Microsoft-only licensees run 56% below unlicensed publishers on Copilot.
Sign with OpenAI and ChatGPT becomes your dominant citation source. OpenAI-licensed publishers draw 57.9% of all their AI citations from ChatGPT alone. Publishers with no deal draw 32.4% and lean on Perplexity almost equally.
58.8% of licensed publishers’ citations come from commercial evergreen content, not journalism. Dated news articles are just 5.1%.
Five media groups control 69% of every citation going to licensed publishers. The licensed pool is far narrower than 314 domains suggests.
Niche and trade media are the biggest winners of AI search. Trade outlets carry the majority of news citations in 15 of 16 industries, collecting 213% more citations than mainstream media, and they are almost entirely unlicensed.
Why Are AI Content Licensing DealsCritical for AI Search Optimization?
Licensing deals are negotiated between AI companies and publishers. PR and marketing teams get no vote in them. So why should you care?
Because they quietly reshape your earned media strategy. If licensed publishers are systematically favoured by an AI platform, then a placement in a licensed outlet is worth more than an identical placement elsewhere, and your pitch list should reflect that. If they are not favoured, then paying a premium for tier-one licensed mastheads buys you prestige and nothing else in AI answers.
Until now nobody could answer that with data, because answering it needs two things at once: a complete, sourced list of who signed what, and citation data at enough scale to measure the effect. Press Ranger built the first. OtterlyAI has the second.
Scope of Study
We joined two first-party datasets.
The citation side (OtterlyAI):
129.3 million citations captured during June 2026
19,812,878 cited URL records across 7 AI Search platforms
ChatGPT, Google AI Overviews, Google AI Mode, Perplexity, Microsoft Copilot, Gemini, Claude
The deal side (Press Ranger AI-Publisher Licensing Research):
91 confirmed AI company x publisher agreements, public evidence through 28 July 2026
Mapped to 314 publisher domains
Every record backed by a public source
90.8% of the citations we attribute to licensed publishers sit on domains explicitly named in deal announcements, not inferred from portfolio ownership. Per-page comparisons divide each group’s citations by its number of cited news pages, so a large publisher does not distort the result simply by publishing more. This is a descriptive one-month snapshot: it shows association, not causation.
The ChatGPT Effect:48% More Citations Per Page
Take every news page cited by AI search in June 2026. Pages from publishers with an OpenAI licensing deal collected 10.2 citations each on average. Pages from publishers with no deal collected 6.9. That is a 48% citation premium per cited page on ChatGPT, and 46% across all 7 platforms combined.
The chart below runs the same test for every platform against its own owner’s deals. Does a deal with the company behind the platform earn you more citations on that platform?
Two platforms say yes. ChatGPT at +48% and Microsoft Copilot at +85%. The other four say no: Perplexity sits at parity, and all three Alphabet platforms cite licensed publishers slightly less than publishers with no deal.
Anthropic is the outlier with no data at all, because Anthropic has not disclosed a single publisher licensing deal. Claude still cites more news than any other platform, which tells you news citations do not require a contract.
Why the Copilot number is not what it looks like
Copilot’s +85% is the biggest bar on the chart, and it is the one worth taking apart, because it is not evidence that Microsoft rewards its own licensees.
Copilot lifts licensed publishers as a class, whoever licensed them. Publishers signed by OpenAI, Amazon, Meta and Perplexity all earn between 61% and 85% more per page on Copilot than publishers with no deal. What Copilot rewards is publisher scale, and the licensed pool is full of large publishers that signed with several AI companies at once.
Three facts settle it:
Microsoft has no exclusive publishers. All 144 of its licensed domains also hold a deal with another AI company, 94 of them with OpenAI.
99% of the Copilot citations reaching Microsoft licensees go to that OpenAI overlap.
Strip the overlap out and the effect inverts. Microsoft licensees without an OpenAI deal average 2.8 citations per page on Copilot against 6.3 for publishers with no deal, which is 56% worse.
The cleanest test is to look only at publishers that signed with a single AI company. Publishers that signed with OpenAI and nobody else earn 112% more citations per page on ChatGPT than publishers with no deal, well above the 45% premium those same publishers average across the other platforms. Publishers that signed only with Perplexity beat their own away-platform baseline by 6 percentage points. Publishers that signed only with Google beat theirs by nothing at all.
OpenAI is the only licensor whose deal shows up as a genuine home-platform advantage.
Sign With OpenAI andChatGPT Becomes Your Whole Mix
The premium is only half the story. The other half is concentration.
Publishers who signed with OpenAI now draw 57.9% of their entire AI search citation volume from ChatGPT alone. Widen the lens to publishers holding any licensing deal and the ChatGPT majority still holds at 55.0%.
Publishers with no deal look completely different: 32.4% from ChatGPT, with Perplexity almost level at 30.7%, and a far more even spread across the rest.
What this means for your GEO strategy: coverage in licensed publishers concentrates your brand’s AI exposure into one platform. That is excellent if ChatGPT is where your buyers are, and a single point of failure if it is not. Coverage in strong unlicensed publishers spreads your exposure across ChatGPT, Perplexity and Google AI Overviews more evenly.
Flip to the platform’s perspective and the same pattern appears from the other side.
On ChatGPT, 22.8% of all news citations go to OpenAI-licensed publishers, while those publishers account for only 17.3% of the news pages ChatGPT cites. They are over-represented relative to their own footprint. No other platform hands its owner’s licensees more than 6.7%.
Licensing Deals MonetizeBuying Guides, Not Journalism
Here is the finding that surprised us most. We classified every cited page by URL pattern to see what kind of content AI search actually pulls from licensed publishers.
58.8% of licensed publishers’ citations come from commercial evergreen content: best-of lists at 38.1%, buying guides and money advice at 15.1%, product reviews and tests at 5.6%.
Dated news articles account for 5.1%. Industry trends and analysis, 2.0%. Risk and due diligence, 1.7%. The categories that look most like classic journalism are the smallest on the board.
The most-cited licensed URLs of the month make it concrete: software reviews on TechRadar, insurance buying guides on Forbes Advisor, health explainers on Verywell Health. All evergreen. All commercial.
What this means for your GEO strategy: when you pitch a licensed publisher, pitch the service desk, not the news desk. A mention inside a buying guide or a best-of list is worth multiples of a mention in a dated news article, and it keeps earning for months.
And a note for PR teams specifically: press releases and wire copy register at under 1% for both groups. AI search is not reading your wire distribution.
Five Media Groups Own69% of the Licensed Pool
314 licensed domains sounds like a broad market. It is not.
Future plc titles take 15.8% of every citation going to licensed publishers. Forbes alone takes 15.7%. People Inc. 15.4%, Conde Nast 11.7%, Hearst 10.7%. Five groups, 69% of the pool.
The market is geographically narrow too: 88.7% of licensed-publisher citations go to English-language publications and 73.0% to North American ones. DACH is the only region we measured where Perplexity, not ChatGPT, leads.
Deal economics stay mostly private. Of 91 confirmed deals, only two have publicly reported values: Meta with News Corp at up to 50 million US dollars per year, and Amazon with The New York Times at 20 to 25 million.
Announcements peaked in the second half of 2024 with 32 deals and have kept coming since. The deal type has shifted too, from one-off archive training licences toward real-time grounding, product distribution and usage-linked models.
The Bigger Opportunity:Trade Media Is Winning AI Search
We joined a second dataset to test this at industry level: 16 US industry prompt sets tracked across the same 7 platforms, cross-referenced against the same 91 deals.
Niche and trade media beat mainstream media in 15 of 16 industries. Finance and Insurance is the single exception, where mainstream outlets take 52.2%. Everywhere else trade press carries the industry’s news presence in AI answers, rising to 92.6% in Nonprofit and 93.3% in Industrial and Manufacturing.
Pooled across all 16 industries, niche and trade media collect 213% more AI citations than mainstream media, 75.8% against 24.2%.
Now split each media class by licensing status, and the strategic picture snaps into focus. Within mainstream media, publishers with licensing deals capture 54.7% of citations. Within niche and trade media, only 17.2%.
AI companies licensed the mainstream layer. The trade layer that carries most industry news is almost entirely unlicensed, and it gets cited anyway.
Every single one of the 7 platforms prefers trade media, from 69.5% on Google AI Mode to 82.2% on Gemini. This is not one engine’s quirk. It is how AI search works.
News Is Only7.2% of AI Answers
One piece of context before you rebuild your media plan around this.
Across 129.3 million citations, News and Media content receives 7.2%. The rest goes to brand pages, blogs, government sites, forums, encyclopedias and video. Claude cites news most at 14.0%, roughly 1.7 times ChatGPT’s 8.3%.
Earned media matters in AI search, but your own site is still the larger prize. Anyone telling you PR is the whole GEO answer is selling you something.
The Licensing Deals Playbookfor AI Search
1. Match your pitch list to the platform you need
Coverage in licensed publishers tilts your exposure toward ChatGPT, where those publishers earn a majority of their citations. If you need Perplexity or Google AI Overviews visibility, strong unlicensed vertical publishers spread you more evenly. Check which publishers already earn citations for your prompts in your OtterlyAI Citations report before you decide.
2. Pitch the formats AI actually cites
Best-of lists, buying guides and product reviews drive 58.8% of licensed publishers’ citations. Ask for inclusion in the roundup, the comparison, the buyer’s guide. A dated news article is the weakest placement you can win.
3. Stop over-indexing on mainstream media
Trade outlets take the majority of news citations in 15 of 16 industries, and they are cheaper and easier to land than national press. Brands spent two decades optimizing for tier-one mastheads. In AI search that is now the more expensive route to less visibility.
4. Do not pay a premium for a licensing badge
Outside ChatGPT, a licensing deal does not correlate with more citations. If a publisher’s rate card assumes their AI deal makes them more valuable, the data does not support that on 5 of 7 platforms.
5. Remember that a deal is not a prerequisite
The most-cited news domains in our data include completely unlicensed outlets: NerdWallet, Finanztip, Healthline, Handelsblatt and Bankrate each collected more citations in June 2026 than every licensed publisher except Forbes and TechRadar. They earned that position with content, not contracts.
Final Conclusion
Our analysis of 129.3 million citations against 91 confirmed AI content licensing deals points to five things you can act on:
OpenAI is the exception, not the rule. A deal with OpenAI correlates with 48% more citations per page on ChatGPT. Deals with Google, Perplexity and Microsoft show no home-platform advantage once you control for publisher overlap.
A deal narrows your platform mix. Licensed publishers draw 55 to 58% of their citations from ChatGPT alone. That is concentration, not just upside.
The money follows service journalism. 58.8% of licensed citations come from commercial evergreen content and only 5.1% from dated news. Licensing deals monetize buying guides.
The licensed market is small and concentrated. Five media groups hold 69% of the pool, and 73% of it is North American.
Trade media is the real opportunity. It wins 15 of 16 industries, collects 213% more citations than mainstream media, and is almost entirely unlicensed. That is underpriced attention.
Explore every chart, filter all 314 licensed publishers by AI firm and platform, and download the full whitepaper in the interactive version of this study.
All figures are first-party OtterlyAI citation data (June 2026, 7 AI Search platforms) joined with Press Ranger’s AI-Publisher Licensing Research (91 confirmed deals, public evidence through 28 July 2026). Industry findings cover 8 to 14 Aug 2026. Descriptive data: association, not causation.
Got a GEO experiment idea?Let’s collaborate
OtterlyAI is a leading AI Search monitoring and optimization platform, trusted by 20,000+ marketing pros. And we don’t write about AI Search from the sidelines. We test it. With the community, we run and update a public GEO Experimentation Tracker showing what we are running and what moved AI Search. If you are running your own GEO experiment, email rick.tousseyn@otterly.ai and we may feature it with your name attributed.
👉 Want to measure your AI Visibility? Start with OtterlyAI and our GEO Guide.
PressRanger is an AI-native PR platform that helps teams get visible in AI search. PressRanger also built the deal research behind this study: the AI-Publisher Licensing Research, 91 confirmed agreements mapped to 314 publisher domains, every record backed by a public source. If the trade-media findings above changed your pitch list, PressRanger is where you build the new one.
👉 Start running performance PR campaigns with PressRanger.
👉 Want the licensing dataset behind this study? Reach out via pressranger.com.
