On July 21, 2026, the Wall Street Journal published a piece that's being called a turning point in how the web works. The headline: Reddit, USA Today, Politico, Reuters, The Economist, and People Inc. are all actively evaluating whether to cut Google off. Not just block the AI training bot. Not just put up a paywall. Block Google's crawler entirely โ which means disappearing from search results altogether. A few years ago, this would have been career-ending for whoever suggested it. Now it's a real strategy discussion happening at the highest levels of some of the largest publishers on the internet.
Let's talk about why, what the numbers actually show, and what any of this means for you if you're trying to get traffic from Google in 2026.
The Numbers That Broke the Deal
The WSJ piece, citing Semrush data, laid out the traffic carnage in black and white. Between June 2025 and June 2026:
- USA Today (national site): organic US Google traffic dropped by nearly 50%
- Business Insider: down more than 85%
- Politico: down 23%
- CNN: down roughly 25%
Those aren't rounding errors. Those are existential numbers for advertising-dependent publishers. When half your search referrals evaporate in twelve months, you start doing math about whether Google is a partner or a competitor using your content as free training data.
What's Actually Driving the Decline
AI Overviews are the main culprit, and the click-through data backs it up hard. When a query triggers an AI Overview, the click-through rate to publisher content drops from around 30% all the way down to about 8%. That's a roughly 73% reduction in clicks โ on queries where Google is surfacing your content as a source but sending almost nobody to actually read it. You're feeding the machine. The machine isn't feeding you back.
Reddit is in an especially awkward position here. The platform signed a $60 million per year deal with Google to let Google use Reddit posts to train its AI models. That deal is up for renewal, and Reddit's executives are reportedly asking a very reasonable question: if Google is using our content to give AI-generated answers that prevent users from visiting Reddit, what exactly are we getting out of this arrangement?
The Impossible Choice Google Built
Here's the part that should make every site owner pay attention. Publishers technically have ways to limit how Google uses their content. You can block the Gemini training crawler via Google-Extended in your robots.txt. But there's a massive catch.
If you want your content excluded from AI Overviews โ the actual thing eating your clicks โ you can't do that cleanly without also tanking your visibility in regular search results. Google's crawler serves double duty: it indexes your content for traditional search and feeds it into AI answer generation. There's no surgical opt-out. It's all or nothing.
Google internally described building a separate AI Overviews opt-out as a "substantial engineering challenge" back in February 2026. Which reads a lot like: we know this is a problem but we're not rushing to fix it because the current setup benefits us.
"No publisher wants to make do without Google, but many are increasingly prepared to do so. Google, on the other hand, has no such backup plan." โ Mark Stenberg, Adweek
The UK's Competition and Markets Authority stepped in last month and ordered Google to create a proper opt-out โ letting publishers exclude their content from AI features without losing search visibility. That ruling only covers UK publishers directly, but it sets a precedent that regulators elsewhere are watching closely.
If you're running a content site and thinking about blocking AI crawlers, be careful. A study of publishers who blocked generative AI crawlers via robots.txt found a persistent 23% decline in monthly visits โ not because Google penalized them, but because being excluded from AI answers reduced brand exposure that was quietly driving indirect traffic. There's no clean exit here. That's the whole point.
What Publishers Are Actually Considering
The options on the table vary by publisher. Nobody's actually pulled the plug yet, but here's where conversations stand as of late July 2026:
USA Today / Dotdash Meredith
CEO Neil Vogel told the Journal that "turning Google off and blocking them entirely is 100% on the table." The company is suing Google over ad-tech monopoly practices and has reportedly started planning for a future where Google traffic isn't a significant revenue driver. That's a publisher with real scale actively de-risking from search.
The $60M/year deal is up for renewal and Reddit leadership is doing a hard-nosed assessment of whether continued Google access makes business sense. Reddit's content is uniquely valuable for training conversational AI โ the kind of informal, experience-based answers that Google desperately wants in its AI products. Reddit knows this. So does Google. That negotiation is going to be interesting.
The Economist
Their VP of generative AI put it plainly: "Traffic can be choppy and declining, but traffic is still traffic." Translation: they're not blocking Google yet. But they're having the conversation. Their focus is on opting out of AI features specifically without losing search presence โ the very thing the UK regulator just forced Google to allow.
Politico
Axel Springer, Politico's parent, is exploring putting registration walls on non-paywalled content โ forcing Google's bot to either authenticate as a human (which it can't) or crawl nothing. That's a creative middle path that stops short of full blockage but limits what Google can scrape.
Why This Matters for Regular SEO Practitioners
You might be thinking: okay, this is a big-publisher problem. I'm not the Economist. Why should I care?
A few reasons.
The Precedent Effect
If major publishers successfully opt out of Google and survive โ or even thrive by forcing users to come directly โ it changes the negotiating dynamic for everyone. It shows that the web doesn't have to feed Google. That's a psychological shift with real downstream effects on how Google treats content creators.
Index Quality Could Drop Fast
If premium, trustworthy publishers pull their content from Google's index, what fills that vacuum? Lower-quality content, spam, AI-generated junk. Some SEOs are already seeing this in certain verticals. Google's index quality problem gets dramatically worse if the publishers it relies on for authoritative signals all leave at once.
Your Traffic Diversification Window Is Right Now
The real lesson here isn't "block Google" โ it's that any business built entirely on Google referrals is fragile. These publishers built massive audiences on search traffic over 20 years. Now that traffic is down 30-85% and they have no alternative channels to fall back on. That's the actual cautionary tale.
The smart move for site owners right now: treat Google as one channel in a diversified mix. Build your email list. Optimize for Google Discover (which is still sending real clicks). Get into YouTube, since it's the most-cited source in AI Overviews. Build brand recognition so people search for you specifically, not just the topic you cover.
What Google Needs to Do (But Probably Won't Fast Enough)
Google's position is essentially: "You can opt out of Gemini training, we send you valuable traffic, this is a fair deal." But the data no longer supports that framing. The traffic Google sends has declined dramatically, and the content publishers create is now the primary fuel for AI products that further reduce that traffic. That's not a partnership. That's extraction.
For Google to stabilize this situation, they need to do a few things:
- Build a real, granular opt-out for AI Overviews that doesn't affect traditional search rankings
- Create a revenue-sharing model for publishers whose content is cited in AI answers
- Be transparent about how much traffic AI features actually suppress versus generate
Some of these are technically complex. Some are commercially unattractive for Google. The UK regulator is forcing the first one. The others are likely to require either more regulatory pressure or publishers actually following through on the threat to leave.
Bottom Line
The publisher revolt happening right now is a symptom, not the disease. The disease is that Google built a system where it profits from content it didn't create, without a sustainable mechanism for compensating the people who create it. That's worked for 25 years because search traffic was so valuable that publishers happily paid with their content. Now that AI Overviews have broken the traffic equation, the math doesn't work anymore.
Watch this story closely over the next few months. If even one major publisher โ especially Reddit โ actually follows through on blocking Google, it will accelerate conversations that have been theoretical until now. And the SEO implications will be significant: a more fragmented index, faster-growing importance of direct traffic and brand search, and probably more regulatory action in the EU and US.
In the meantime, do the boring but real work: diversify your traffic sources, build your brand, make content that people actually seek out by name. Don't be completely dependent on any single platform โ including Google.
๐งช Check Your Site's SEO Health
Run a free 60-factor audit on your site. No login, no credit card. Just results.
Run Free SEO Audit โ