Anthropic and the Associated Press struck a content and data licensing deal just before the US midterm elections, Adweek reported on 4 November, citing two people familiar with the matter. Adweek says the arrangement marks the first time Anthropic has paid a digital news publisher to license its content.
The contrast Adweek draws is with OpenAI, which it says “has negotiated more than a dozen content licensing deals with major publishers”. By our count of our own coverage, Anthropic is the fourth AI developer whose licensing deal with news publishers we have reported, after Amazon, OpenAI and Google.
Business Pill · Distribution Advantage
A short Business Engineer explainer on distribution advantage. It teaches the general idea only and says nothing about Anthropic, the AP or this deal.
The key insight: Adweek reports that Anthropic’s first paid licence with a digital news publisher is with the AP. By our count of our own coverage, that makes Anthropic the fourth AI developer with a news-licensing deal, and Adweek’s report came 161 days after the AP named OpenAI an elections customer, by our arithmetic.
Mark Stenberg wrote in Adweek that a Claude user asking for the latest polling data on election night would have seen “up-the-minute information, provided by and attributed to the Associated Press”. That, the report says, was the result of the licensing deal.
The terms are not public. Adweek wrote: “The financial specifics of the arrangement, as well as other critical details, such as its timeframe, were not disclosed by either party.” It says Anthropic did not respond to a request for comment and the AP did not provide a comment on the record.
By 01:10 UTC on 5 November we had found no announcement on Anthropic’s news page or AP’s press release page, and no post on the deal in the latest posts we read from the AP, Anthropic, Adweek or Stenberg on X.
Publisher licensing deals as reported by Adweek on 4 November 2026: OpenAI “more than a dozen”; Anthropic one, with the AP, according to two people familiar with the matter. Not independently verified; terms not disclosed.
Why Anthropic Moved, in Adweek’s Account
Adweek says Anthropic had resisted publisher deals. In August, digital media executives told Stenberg that the company believed its scraping and use of publisher content was legally permissible under fair use.
Stenberg’s explanation for the change is his own analysis: “an information index without access to up-to-date data is dramatically limited in its utility.” He writes that publishers had been blocking Anthropic’s crawlers, a tactic he calls “effectively a digital embargo”, and that “in this limited instance, the strategy worked.”
He also hedges the scope. Adweek wrote that the deal “is by no means an indicator that Anthropic has irrevocably changed its licensing strategy.”
The Election-Night Link
On 14 October Anthropic said that “Claude will begin showing up-to-date results on election night and in the days that follow”, so that people asking for vote counts would see “the latest official numbers directly in their conversation.” The passage we read named Democracy Works as the source of voting information but did not name a source for the results.
The AP sells exactly that kind of feed. On 27 May the AP said it would provide OpenAI with “vote count results data for national, state and local races” through the 2028 general election. By our arithmetic, Adweek’s report on Anthropic came 161 days after that announcement. Adweek does not say whether Anthropic’s deal covers AP’s vote count and race calls, or other AP content.
For the midterms, the AP said it would count the vote and declare winners in nearly 7,000 races. Adweek also notes that the AP “was also the first publisher with whom Google and OpenAI signed licensing agreements.”
Who Pays for News in AI
The deal lands a day after a Brazilian court went the other way on Meta. A São Paulo state court ordered Meta on 3 November to stop collecting content from Brazilian media outlets in three national press associations to train its AI models, AFP reported, with a daily fine of R$1 million capped at R$100 million.
The other deals in our count: Amazon is paying The New York Times between $20 million and $25 million a year, according to reports in July 2025. OpenAI signed ChatGPT content partnerships with two Indian publisher groups in September, with no fee disclosed. Google said on 15 October that more than 400 publications take part in its News AI pilot program, which pays them. Our count covers our own reporting only and is not a census of the market.
The business lever is the answer engine’s input cost. A licensed feed turns news from content a model crawls into a supplier contract with a price, even if that price is undisclosed.
As we read the report, election night tested whether Claude could answer a live question that the unlocked web could not, and the AP was the supplier that could. Adweek frames the deal as a test of licensing, and says Anthropic is preparing to go public in a matter of weeks.
Connecting the Coverage
Our related pieces on the midterms, the AP’s race calls and Anthropic:
As we read Anthropic’s 14 October post next to Adweek’s report, the election-night results Claude promised needed a supplier, and Adweek says the information users saw that night was attributed to the AP.
As we read the AP’s releases, the AP already supplies election results to technology platforms, including OpenAI through the 2028 general election. Adweek does not say whether Anthropic’s terms resemble those.
As we read the Brazil order and this deal side by side, news publishers are pressing AI companies through courts and through contracts in the same week.
Mark Stenberg, Adweek, 4 November 2026
“an information index without access to up-to-date data is dramatically limited in its utility”
Three Implications
FOR ANTHROPIC Adweek says the company is preparing to go public in a matter of weeks; the deal it reports adds a supplier contract with a news publisher, with no disclosed price or duration.
FOR THE AP The AP says the world’s leading media, technology and financial companies rely on it for election news and data, and Adweek says Google and OpenAI each signed their first publisher licensing agreement with it.
FOR PUBLISHERS Adweek describes crawler blocking as a tactic that worked in this limited instance; our earlier pieces cover the other route, the courts, from Brazil to The New York Times v. Brave.
The framework describes the Corpus Moat as “built by pretraining, eroding as the public web is exhausted and synthetic data closes gaps.”
As we read it, a licensed AP feed sits on the other side of that line: a live input supplied under contract, where the corpus was gathered from the open web. The framework does not discuss the AP or this deal.
What Is Not Established
Neither company has confirmed the deal on the record in anything we found. The price, duration and scope are not public, and no source we read says whether the deal covers AP’s archive, its race calls or its news stories beyond election coverage.
The “more than a dozen” OpenAI figure is Adweek’s. The fourth-developer count and the 161 days are ours, from our coverage and AP’s release date.
This Week In The AI Supercycle
Anthropic’s Reality Check, The Capital Squeeze, & The New World Order
This week’s issue: for two years the AI build-out was priced as a growthstory. This week it started being priced as a credit story.
Anthropic has paid a digital news publisher to license content for the first time, in a deal with the AP struck before the midterms, Adweek reported, citing two people familiar with the matter. The terms were not disclosed, and neither company commented on the record. By our count, Anthropic is the fourth AI developer whose news-licensing deal we have reported.
A note on sourcing. The deal, the first-paid-publisher description and the OpenAI comparison are from Adweek’s report of 4 November 2026, which cites two people familiar with the matter; neither company commented on the record. Election data details are from AP’s releases of 27 May and 28 October 2026 and Anthropic’s post of 14 October 2026. The 161 days is our arithmetic; the fourth-developer figure is our count of our own coverage.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
