OpenAI Publisher Partnerships 2023–2026: Complete Roadmap of Every Major Publishing Deal

OpenAI's agreements with publishers generally encompass one or both of two aspects:

Training data: permission to utilize a publisher's archive for model training, encompassing paywalled material.

Live display: the ability to showcase summaries, quotes, and links from the source within ChatGPT, particularly in ChatGPT search, with proper credit.

Numerous agreements include a third component: the publisher gains access to OpenAI technology to create its own products. OpenAI obtains high-quality, legally compliant content and up-to-date news for its responses. Publishers receive income, recognition, and representation as courts determine if AI's use of copyrighted material constitutes fair use.

All OpenAI publisher Partnerships agreements listed below were confirmed by cross-referencing the initial announcements and unbiased media reports. Financial figures are sourced from journalists and have not been verified by the companies.

2023: Templates

Associated Press (July 2023). The first significant news agreement in the AI age. AP granted a license for its text archive dating back to 1985 for training and received access to OpenAI technology in exchange. The deal allegedly featured a first-mover protection enabling AP to revisit negotiations if a subsequent publisher secured better conditions, an initial indication that publishers anticipated price increases.

Axel Springer (December 2023). The large German company that owns Politico, Business Insider, Bild, and Welt established the initial partnership blending training rights with real-time, attributed news summaries within ChatGPT, encompassing paywalled material. Bloomberg stated the worth at several tens of millions of euros over approximately three years. This training-and-attribution model established the blueprint for almost all subsequent agreements.

2024: Land Acquisition

Le Monde et PRISA Media (March 2024). OpenAI's initial non-English initiative, revealed together. Le Monde delivered France’s leading newspaper. PRISA provided Spain’s El Pais, the business newspaper Cinco Dias, the sports daily AS, and El Huffpost, allowing ChatGPT to access licensed Spanish-language news for Spain and Latin America. Le Monde subsequently revealed that approximately 25% of its OpenAI earnings are distributed to its journalists' association, providing one of the rare insights into the financial arrangements.

Financial Times (April 2024). An agreement that is non-exclusive, encompassing training along with attributed quotes and summaries that include links to FT.com. At that time, OpenAI stated it had approximately a dozen agreements either signed or forthcoming.

Dotdash Meredith (May 2024). The publisher owned by IAC (renamed People Inc. in 2025) has licensed over 40 lifestyle brands such as People and Investopedia. This agreement is significant as it yielded one of the few concrete public figures: IAC’s reports disclosed that OpenAI pays a minimum of $16 million annually in fixed charges along with a variable portion. OpenAI models also drive the advertisement-targeting product of the publisher.

News Corporation (May 2024). The standard for the industry. The Wall Street Journal, Barron’s, MarketWatch, the New York Post, The Times of London, The Sun, and News Corp’s Australian publications, with the agreement said to be valued at over $250 million in cash and OpenAI credits over five years, according to the WSJ’s own reporting. HarperCollins, which is also owned by News Corp, was excluded.

Vox Media and The Atlantic (May 2024). Vox Media acquired The Verge, Eater, and New York Magazine; The Atlantic obtained a license for an archive dating back to 1857. Both agreements sparked public protests from newsroom unions, a consistent issue throughout the licensing surge.

TIME (June 2024). A long-term agreement spanning 101 years of archives, including attribution and links in ChatGPT alongside OpenAI technology for TIME's own offerings.

Conde Nast (August 2024). Vogue, The New Yorker, Wired, GQ, and Vanity Fair collaborated with ChatGPT and the SearchGPT prototype. Significant because CEO Roger Lynch was among the most vocal advocates in Congress calling for AI compensation prior to signing.

GEDI (September 2024): signed, but then paused. The Italian publisher of La Repubblica and La Stampa signed; however, Italy’s data protection authority released an official GDPR warning in November 2024. GEDI announced that the project had not commenced and no material had been distributed. It remains the most evident instance of European rules delaying an AI licensing agreement.

Hearst (October 2024). Over 20 magazines and 40 newspapers, such as Cosmopolitan, ELLE, Esquire, the Houston Chronicle, and the San Francisco Chronicle, appeared in ChatGPT along with citations and links.

Future plc (December 2024). More than 200 expert brands such as TechRadar, Tom's Guide, and PC Gamer. Notably, Future informed investors that the partnership held no financial significance, implying that its worth lay in strategic visibility instead of monetary gain. Future has since embraced its prominent presence in ChatGPT as a business advantage.

2025 to 2026: Financing Newsrooms and Resolving Legal Disputes

Axios (January 2025). In a first, OpenAI directly supports journalism by funding four new Axios Local newsrooms (Pittsburgh, Kansas City, Boulder, Huntsville) for three years, in addition to content licensing.

Schibsted Media (February 2025). The initial Nordic agreement: Norway’s VG and Aftenposten, along with Sweden’s Aftonbladet and Svenska Dagbladet, offer cited news summaries in ChatGPT.

The Guardian (February 2025). A significant reversal from one of the most openly doubtful publishers. The agreement includes attributed summaries and excerpts, along with an internal deployment of ChatGPT Enterprise by the Guardian.

The Washington Post (April 2025). A partnership centered on display that integrates summaries, quotes, and links into ChatGPT searches. During this announcement, OpenAI stated it had over 20 publisher partners that encompassed over 160 outlets in more than 20 languages.

Grupo Folha and Grupo UOL (May 2026). OpenAI’s initial agreements in Brazil, marking the first definite conversion of a lawsuit into a license: the accord resolved Folha de S.Paulo’s copyright lawsuit against OpenAI for 2025. Brazil ranks among ChatGPT’s top markets, boasting more than 50 million active users each month. The publishers also gain access to Codex, ChatGPT Enterprise, and API.

A notable pattern is that no new OpenAI publisher partnership was confirmed between the Washington Post agreement in April 2025 and the Brazil agreements in May 2026.

The latter part of 2025 was dominated by agreements involving Microsoft, Amazon, and Meta instead. Approach with doubt any list asserting OpenAI partnerships during that period.

Aside from news publishers, OpenAI secured three major platform data agreements:

Shutterstock (July 2023): a six-year license encompassing images, video, music, and metadata for training, building on a partnership that began in 2021.

Stack Overflow (May 2024): access to OverflowAPI integrates verified developer Q&A within ChatGPT with acknowledgment, enhancing OpenAI models for programming.

Reddit (May 2024): immediate access to Reddit information through its data API, with an estimated worth of approximately $70 million annually (not officially verified). Reddit material is currently one of the most referenced resources among AI assistants.

For each publisher that obtained a license, others pursued legal action. The two routes are closely linked: each decision alters the negotiating power for the subsequent collaboration.

The GEMA verdict in Munich is noteworthy for those monitoring AI regulation in Europe: it was the first European merits ruling stating that content stored in model weights can violate copyright, with fines reaching 250,000 euros for each infraction. The Folha agreement demonstrates an alternative approach: using litigation as a means to negotiate a license.

Numerous widely discussed negotiations failed to yield an announcement, and they should not be referenced as agreements:

CNN and Fox Corp: discussions on licensing noted in January 2024, featuring video; no announcements made since.

Indian publishers: no licenses from OpenAI are available in India; a statutory royalty scheme was proposed by India in late 2025.

AFP: the French news organization authorized for Mistral, not OpenAI.

Japanese and Korean publishers: no confirmed OpenAI content agreements; Japanese publishers are suing Perplexity, and OpenAI’s partnership with Kakao in Korea is a product arrangement, not a content license.

These agreements go beyond news in the media sector. They directly influence what ChatGPT says and references:

Content from licensed publishers can be displayed. A licensing agreement may provide summaries, citations, and links, but it doesn't ensure a citation or preferred positioning. Coverage in El Pais, the Financial Times, TechRadar, or People now includes distribution in AI responses.

The pool of citations is merging. With over 160 licensed outlets on one side and publishers facing blocks or legal battles on the other, the range of sources AI assistants rely on is changing. The extent of coverage for your brand significantly influences whether AI assistants refer to it.

Digital PR features a new measurement standard. A characteristic in a licensed publication may serve as one of the sources an AI assistant accesses, but citation is not assured. This turns publisher selection into a tactical GEO choice, rather than merely a choice of reach.

The essential inquiry is assessment: which references do ChatGPT, Perplexity, Gemini, Google AI Mode, and Google AI Overviews genuinely cite when discussing your category, and is your brand included in those responses?

AI monitoring tools monitor mentions, citations, and share of voice for your brand across all five default models, provides sentiment analysis, indicates which publishers and domains influence AI responses in your market, and translates that into actionable optimization with recommendations. If a licensed publisher controls the citations in your category, you will notice it, and you can respond accordingly.

OpenAI Finds Publishers with Valuable Content

Publishers that generate authoritative, exclusive and frequently updated information on diverse themes such as news, business, technology, finance, and lifestyles, among others, are the main partners of OpenAI.

This is done in order to enrich OpenAI's database with better, accurate and diverse information. Publishers with good editorial standards and exclusive reporting services are usually prioritized because of their exclusive and valuable content compared to syndication and poor-quality information.

• Prioritizes original reporting publishers with good editorial standards.
• Values not only archives but also news updates regularly conducted by the publishers.
• Enter into licensing deals with publishers rather than relying on only publicly available web information.
• Values information that will increase the performance of the models.

Content Rights are Negotiated by Both Parties

Every agreement with publishers is negotiated individually, thus, there is no universal contract that is applicable to all partners. It is determined how OpenAI is allowed to use articles for training models, showing summaries and extracts in ChatGPT Search, or achieving both goals at once.

Contracts determine attribution requirements, limitations on using content, terms of license, and commercial terms. Not only content licensing but also access to OpenAI technology like ChatGPT Enterprise, API, or other AI-based tools.

• Agreements determine whether content will be used for training, displaying or for both.
• License terms include attribution rules, content usage terms, term of license, and payment terms.
• The owner of the content remains the same, however, certain permissions are granted to OpenAI.
• In many cases, there is an opportunity to access products of OpenAI.

Licensed Content Gets Integrated Into the OpenAI Ecosystem

After the completion of the process, the content licensed by the publishers gets used by OpenAI depending upon the rights mentioned in the license. The training license helps future AI models learn from the authorized archives, whereas the display license allows ChatGPT to show summaries, excerpts, and references.

Licensing does not ensure that the published material gets used in all responses. It is up to ChatGPT to decide whether licensed material is cited or not, depending upon its relevancy to the question asked by users.

• Training licenses help future AI models perform well from approved publisher archives.
• The display licenses help ChatGPT to surface summaries, excerpts, and references with attribution.
• Licensed content gets used but does not ensure its inclusion in the AI responses.
• The citation depends upon relevancy and context of the user's question.

Publishers Earn Financial and Strategic Gains

These partnerships bring more than just a licensing fee to publishers. They offer new sources of income and greater reach through one of the world's fastest growing AI platforms. Publishers who have their work featured on ChatGPT can enjoy greater brand awareness and referral traffic from the attributed links.

Many of the partnerships even involve access to the enterprise offerings of OpenAI, API credits, technical partnership, or AI-enabled newsroom solutions to enable publishers to enhance editorial, ad tech, customer, and product experiences.

• Licensing deals offer another income stream alongside ads and subscriptions.
• Attributed links can help publishers raise their visibility and draw readers to their sites.
• Publishers often get access to enterprise offerings from OpenAI, API credits, and AI-enabled newsroom solutions.
• Strategic partnerships help publishers explore AI products and optimize operations.

OpenAI has established the most extensive publisher licensing portfolio in AI, comprising around 20 verified partnerships that encompass over 160 outlets in more than 20 languages, highlighted by the reported $250 million collaboration with News Corp and the $16 million annual commitment from Dotdash Meredith.

The approach transitioned from archive licensing (AP) to real-time attributed news (Axel Springer onward), to financing newsrooms (Axios), to resolving lawsuits with agreements (Folha and UOL).

In the meantime, over a dozen publishers opted for legal action, and decisions such as GEMA in Munich are altering the balance of power. For brands, the message is clear: AI responses now feature a licensed supply chain, and visibility within it can be monitored and controlled.