By Peter Wilmot, Chief Product Officer at Shopsense AI
Until recently, the open web operated under a simple, informal arrangement: publishers produced content, Google indexed it, and users discovered that content through search results that directed them back to publisher sites. In turn, publishers monetized that traffic with ads, and Google monetized search queries with their own.
Both sides made money through a status quo that allowed Google to make search useful and publishers to make their content discoverable. The architecture of the web (crawlers, links, rankings) held the whole thing together.
That was until February 12, 2026, when the Penske Media Corporation filed a 56-page federal memorandum arguing that Google broke this deal—not gradually or accidentally, but deliberately. And in a way that undermines the very foundations of monetization on the open web.
What Penske is actually arguing
Given that Penske Media publishes titles including Rolling Stone, Deadline, Variety, Hollywood Reporter, and IndieWire, it’s unsurprising that the company’s filing focuses on the conditions under which publishers have allowed their content to be crawled.
Originally, publishers like these granted Google permission to index their pages, namely because indexing generated traffic. If that traffic had disappeared, they could block Google’s crawlers using robots.txt. The tenuous arrangement lasted, though, largely because both sides captured value.
AI Overviews have changed that flow significantly. Instead of directing users to publisher sites to read the underlying material, Google increasingly aggregates information directly in the search interface. As a result, users often receive the answer they were looking for without needing to leave the results page, which means the publisher receives neither the visit nor the opportunity to monetize it.
Penske’s argument is that Google has effectively altered how publisher content generates value, but that it continues to access that content under the same crawling permissions.
The numbers behind the claim
Traffic data has made this shift visible. Global Google search referrals to publisher sites declined by 33% last year, with U.S. referrals dropping by 38%. Discover traffic has also fallen significantly.
When AI Overviews appear in search results, click-through rates to publisher sites also decline sharply, while zero-click searches (queries where users never leave the results page) continue to climb. This has led publishers like Business Insider to report losing more than half of their organic traffic.
Now, these consequences are leading many publishers to reconsider how they approach search indexing, with a growing share now saying they may block Google from crawling their content.
The walled-garden response
While Penske pursues this issue in court, bigger companies are experimenting with different economic models for publisher content. Amazon Web Services, for example, is developing a marketplace where publishers can license their content to AI developers, with pricing based on usage volume. Microsoft is drumming up their own version, the Publisher Content Marketplace.
Models like these introduce a paradigm quite different from the traditional arrangement. Instead of granting access to content in exchange for traffic, publishers would treat their work as licensable data within the AI ecosystem. Whether these marketplaces generate meaningful revenue remains uncertain, but they do introduce an alternative—even if it’s one that doesn’t quite solve the issues plaguing the open web.
Three fights, one question
The debate over publisher content is unfolding in several arenas. In the courtroom, publishers argue that the original exchange between search platforms and content creators has been altered without consent. In emerging marketplaces, companies like Amazon and Microsoft are experimenting with licensing systems for AI training data. At the same time, Google steams ahead by expanding its AI-generated answers and commerce features in search.
Each of these developments begs the same underlying question: Who will capture the economic value created by digital content when AI systems mediate access to information?
The situation is complicated by the fact that time is not exactly on publishers’ side. Legal disputes take years to resolve, while product changes affect traffic almost immediately. For an example, one need only remember the Department of Justice’s antitrust case against Google from 2020, which remains unresolved. Penske’s lawsuit will likely take just as long to play out. Meanwhile, Google’s AI features will continue to scale and reshape search behavior.
The reality facing publishers is thus as simple as it is dire: the existing bargain is failing before anything resembling a replacement has fully emerged.
Rebuilding the open web’s economy
For those of us who still believe in the open internet, the Penske dispute serves as a warning about where things are headed if we don’t act.
For decades, the open web relied on that simple loop wherein publishers created content, search engines directed users to it, and advertising revenue funded further production. AI-generated answers are weakening that loop by reducing the number of users who visit the original sources of information.
The walled gardens likely won’t feel the same pressure, as platforms such as Meta, Amazon, and TikTok largely operate inside padded ecosystems that meld user data and commerce infrastructure. Instead, it’s the independent outlets and creators who will keep hurting.
If the existing model continues to deteriorate, the answer cannot be to simply hope that the market corrects itself. We need a new bargain, which means building a new model for monetization that will deliver on the open web’s promise.
