Media Briefing: Turning scraped content into paid assets — Amazon and Microsoft build AI marketplaces

Details remain scarce, beyond the fact that publishers provide the supply while AI model developers represent the demand side. But Amazon Web Services reportedly previewed the idea with publishers ahead of its publisher symposium in NYC on Feb. 10. One attendee told Digiday they were expecting to hear more about Amazon’s AI content marketplace plans, but nothing was announced at the event.

Meanwhile, six months after it began its AI content marketplace pilot, Microsoft brought in Yahoo as its first external buy-side partner, having run initial tests with its publisher partners and its own CoPilot demand. Microsoft kicked off its marketplace with Business Insider Inc, Vox Media Inc, USA Today Co., People Inc, The Associated Press, Hearst Magazines and Condé Nast.

Last month, Yahoo launched Scout, its AI search engine and chatbot, which pulls information from Yahoo’s content and other publishers with inline citations and links to sources. Yahoo is a supply-side partner to Microsoft’s marketplace, and can also license other publisher content to feed into Scout — all of which are rights-cleared sources, meaning they are legally permitted to use the content, according to a source familiar with Yahoo’s role in the Microsoft marketplace, who spoke under the condition of anonymity. (And to show how deeply both companies are intertwined, Scout is built using Microsoft’s grounding API.) Publishers get a revenue share of the cost that Yahoo pays for the content in Microsoft’s marketplace, based on a CPM model, according to the source.

Then there is Cloudflare, also pushing toward building an AI content licensing infrastructure for its publisher clients. The CDN vendor recently acquired U.K. startup Human Native to help accelerate its efforts to turn publisher content into AI-ready data that can be sold for licensing.

As always, the same elephant remains in the room: Can these marketplaces compel the LLMs to pay up for content access?

They certainly have a good shot. Microsoft AI vp Nikhil Kolar recently told Digiday that they want to crack open the opportunity for AI models — whether it’s a new AI-developed app, a consumer LLM or a private enterprise model — they will all be able to reach the best and largest information repository (AKA any publisher or website that wants to plug into the marketplace on the supply side) that contains verified, fact-checked data, with no friction. That’s the play.

And it’s already getting early data to show that better results for AI models come from premium publisher partners. “We have seen that without access to premium content versus access to premium content, we see a marked improvement in the customer success, or session success, number of clicks and citations that you get,” Kolar said.

That’s an outcome Human Native co-founder James Smith, has also seen in abundance with its own clients and a large reason why it was snapped up by Cloudflare. Ethics and legality aside, scraping the web is ultimately not good for AI developers’ own products — they just need to be shown a better option, he stressed. 

Publishers welcome the competition  

Mark Howard, chief operating officer at Time, thinks Amazon entering the AI licensing market is positive for publishers. It means competition and more options for publishers to monetize.

“For the good of the open web, we do need more options for how the AI companies can access the information, how they can transact on it. And I do think that — whether it’s [a lot of different] startups… or the big enterprise companies who bring a lot of engineering firepower — it all leads in the right direction,” he said.

Amazon doesn’t have a single flagship model that competes head-to-head with Google’s Gemini or OpenAI’s GPT in broad consumer usage, but it does offer its own lineup of AI models (Titan, Nova) and access to many others (like Anthropic’s Claude) via cloud services (Bedrock.) It also has its AI shopping assistant Rufus, which is still expanding its capabilities.

“Amazon’s entrance here is not significant for stemming the AIO/GPT traffic fears, but it does bring a heavyweight into the ring,” said Scott Messer, founder and principal of Messer Media. 

Publishers want to scale marketplaces so that the buyers (any type of small, large, enterprise, or consumer AI models) are incentivized to pay for content. Microsoft AI’s Kolar recently told us it wants to extend its marketplace to include more domestic and international publishers and AI demand partners.

Some remain wary of any whiff of big-tech dominance. “My concern is that if it [Microsoft] does end up occupying a gatekeeper position in the market, the temptation and shareholder pressure to use that in ways that are damaging to the supply side will be immense,” said David Buttle, founder of DJB Strategies. “With Google, it took place over decades and was, in effect, death-by-a-thousand-cuts. But look at what’s happening now with AI Overviews. It wouldn’t be able to do that if publishers could go elsewhere to distribute their IP,” he added. 

In short, the market needs multiple marketplaces to provide the right level of competition, noted Buttle. 

That’s why Amazon’s push is largely being viewed as good by publishers, so far. For Amazon, launching a marketplace could pay off big if these platforms catch on and AI models start licensing content at scale, noted Messer. “We’ll see if Amazon’s content marketplace is an additive needle-mover for compelling LLMs [to pay publishers],” he said. 

And Amazon could enforce its rules in a totally different way from other marketplaces like Microsoft’s. Take the fact that the existing language in its contracts specifies that its servers won’t run, for example, Pirate Bay or other IP-infringing services. That could potentially extend to AI copyright-infringed content, explained Messer. “Imagine if AWS language said, ‘no using stolen content on our servers’…As for enforcement, the threat of termination should give bootleggers pause before jeopardizing their entire stack,” he said. 

Ultimately, the success of these marketplaces hinges on getting enough demand from the AI developers and LLMs that are taking the content to feed their models.

“Up until now, any AI company has been able to get the data that they want just by going and scraping the open internet. So [AI content marketplaces] have to prove that by transacting this way, there’s value creation on the demand side,” the publishing exec said. “This will come down to demonstrating that these marketplaces make content easier to access and source, and more valuable to AI companies that can provide better quality outputs for users.”

What we’ve heard

“There is a fundamental misconception about the impact of AI on News Corporation. AI is retrospective and synthesizes generic content, sometimes imperfectly. But is past tense, often past imperfect. We have contemporary, creative, proprietary content, which is only accessed if AI companies pay us… We’ve been consciously building a moat, and it is a moat with saltwater crocodiles, with sharks, and an even more dangerous species — lawyers.”

– Robert Thomson, CEO of News Corp, during the company’s latest earnings call on Feb. 5.

Numbers to know

4.3 million: The number of The Wall Street Journal’s digital-only subscribers last quarter, up 13% year over year.

12%: The year-over year growth in digital ad revenue at Dow Jones — which includes publications such as The Wall Street Journal, Barron’s and MarketWatch, among others — according to its latest earnings report.

44-47.5%: The percentage of The Washington Post’s staff getting let go, representing about 350 and 375 people.

30%: The percentage of staff cut at local publisher 6AM City, as it increasingly uses AI to put together its newsletters.

1 million: The number of times ChatGPT is asked about local news in a week.

What we’ve covered

AI licensing deals, protection measures aren’t slowing web scraping

• New data is reinforcing a structural shift in how AI systems access publisher content: AI models are increasingly scraping publisher content, regardless of bot-blocking measures or content licensing deals meant to control usage, improve attribution or drive referral traffic.
• Digiday has pulled four charts that reveal the latest in AI web scraping, and what it means for publishers.

Read more here.

Dow Jones, Business Insider and others on AI-driven search challenges

• In the latest Digiday+ Research report, well over half of publisher respondents said they saw traffic declines last year, and 54% said their traffic decreased somewhat or significantly throughout the year in 2025.
• Nineteen percent of survey respondents each said traffic changes due to social media and AI and AI-driven traffic decreases will be the biggest challenges for the publishing industry this year

Read the report here.

The scale of the challenge facing publishers and politicians eager to damage Google’s adland dominance

• Despite legal challenges, preliminary EU rulings and growing dissatisfaction among advertisers, Alphabet posted record ad revenue in 2025.
• There’s rising pressure to disrupt Google’s grip on ad tech, but these efforts face an uphill battle against the scale and financial strength of Google’s ad operations. 

Read more here.

The Athletic invests in live blogs, video to insulate sports coverage from AI scraping

• As the Super Bowl and the Winter Olympics collided this week, The New York Times-owned publication The Athletic played up coverage that is harder for AI bots to lift: live blogs and video. 
• The Athletic sees live formats as a way to keep audiences on its platform, while insulating its reporting from getting scraped and repurposed by AI tools.

Read more here.

WTF is the IAB’s AI Accountability for Publishers Act (and what happens next)?

• The Interactive Advertising Bureau has proposed new protections to stop AI bots from freely harvesting online content, in a draft bill called the AI Accountability for Publishers Act.
• The proposed legislation seeks to hold AI companies accountable for illicit scraping and failing to comply with publishers’ “no crawling” specifications in their robots.txt files.

Read more here.

What we’re reading

European news publishers accuse Google of using their content in AI Overviews

European news publishers have filed a formal antitrust complaint against Google to the European Commission, alleging that the tech giant is using their content in its AI tools like AI Overviews without proper permission or payment, Reuters reported.

Washington Post CEO Will Lewis steps down

WaPo CEO Will Lewis announced he was stepping down on Saturday, days after the newspaper underwent a massive sweep of layoffs. Jeff D’Onofrio, WaPo’s CFO, was named acting chief executive, The New York Times reported.

How some European publishers are managing Google traffic decline

Publishers – like the Daily Mail, The Telegraph and Bauer Media – see Google search traffic declining largely due to AI-generated summaries and changing user behavior, and are shifting away from pure SEO tactics toward content strategies focused on direct engagement and diversified platforms, Press Gazette reported.

Candle Media sells ATTN: back to its founders

Candle Media has sold a majority stake in digital media company ATTN: back to its co-founders Matthew Segal and Jarrett Moreno, Axios reported. ATTN: was acquired by Candle Media in 2022 for around $150 million.