The Google Network is shrinking and nobody will say what it means

Alphabet's Q1 2026 earnings landed this week and the headline numbers were what they always are now: Google Search revenue up 19% year-on-year to $60.4 billion, queries at "an all-time high," Sundar Pichai sounding relaxed on the call. The story most outlets ran was the usual one. Search is fine. AI Overviews aren't killing the business. Move along.

Look at a different line in the same report and a less comfortable picture appears.

Google Network revenue — the segment covering AdSense, AdMob, and Google Ad Manager, the ads Google brokers on other people's sites — fell below $7 billion for the first time. It's been declining every quarter for two years. In Q1 2024 it was about 12% of Google's total ad revenue. Now it's roughly 9%. And this is happening while the wider US programmatic market grew 20.5% in 2025 to $162.4 billion, according to the IAB/PwC report.

So the programmatic ad economy is booming. Google's slice of revenue from ads served on the open web is shrinking. Those two facts only fit together one way, and the implications for anyone who runs a publisher-adjacent business in the UK are worse than the trade press is letting on.

Two economies, one report

The Alphabet earnings tell a story of two web economies diverging.

The programmatic ad economy is booming. Google's slice of revenue from ads served on the open web is shrinking.

The first is Google's owned surfaces — Search, YouTube, Gmail, Maps, Discover, AI Overviews, AI Mode. Revenue here is accelerating. Pichai's "all-time high" queries comment is almost certainly true. AI features are creating new query types, retail and finance and health are throwing off strong commercial intent, and Google is monetising it all aggressively on its own real estate.

The second is everything else. The independent publisher web. The blogs, niche review sites, vertical media properties, regional news, hobbyist forums, comparison sites — the long tail that built itself around AdSense over the last twenty years. That economy is in measurable decline as a Google revenue source, and the decline is structural rather than cyclical.

The 9% number isn't dramatic on its own. Big companies have segments that wax and wane. What makes it interesting is the direction of travel against a rising market. If programmatic grew 20% and Google Network shrank, Google is either losing share to the rest of the programmatic stack — possible — or the inventory it brokers is itself declining in volume and value. Both can be true at once.

The programmatic ad economy is booming. Google's slice of revenue from ads served on the open web is shrinking.

In January, a two-day technical failure in Google's ad exchange led AdSense publishers to report eCPM and RPM drops of 50–90% with no corresponding traffic decline. Google fixed it. But the episode revealed how thin the margins are for sites whose entire business depends on Google brokering their inventory. When the broker has a bad two days, half your revenue evaporates and there's nothing you can do about it.

That fragility is the actual story.

The publisher base is being cooked from both ends

What's happening to mid-market publishers — and I mean the typical UK independent site, the £200k–£2m turnover business built on display ads and affiliate revenue — is a textbook squeeze.

From the top, AI Overviews and AI Mode are answering more queries on Google's surfaces. Even if you don't accept the strongest version of the click-loss argument (and reasonable people are still arguing about the magnitude), the direction is clear: more sessions resolve on Google, fewer continue to publisher sites. Less traffic means fewer impressions means lower revenue, regardless of CPM trends.

From the bottom, the AdSense economics on the traffic that does arrive are softening. Some of this is Google extracting more margin. Some of it is the slow erosion of ad-supported publishing as a viable model when AI summarisers commodify the content. Some of it is the open programmatic market routing around Google. All of it points the same way for the publisher.

And from the side, a third pressure that gets less coverage: the open web itself is becoming a worse advertising surface as AI-generated slop floods the index. Brands are increasingly nervous about programmatic placements next to low-quality AI content. That nervousness pushes spend toward known-good inventory — Google's owned surfaces, the big walled gardens, premium publishers with editorial controls — and away from the long-tail open web that AdSense was built to monetise.

Three pressures, one direction. The mid-market publisher is being cooked.

Why the SEO industry isn't covering this

Here's a question worth sitting with: why is the trade press so quiet on the Google Network number?

Part of it is structural incentive. The SEO industry sells optimisation services to people who want to rank in Google. Acknowledging that the economics of being a Google-traffic-dependent publisher are deteriorating undermines the pitch. It's easier to talk about Preferred Sources, schema markup, and AI citation tracking than to write the piece that says "the foundational business model that funded most of the content the SEO industry was hired to optimise is breaking."

Part of it is that publisher economics are unsexy. "Programmatic CPMs are softening" doesn't generate the engagement that "Google's new AI feature changes everything" does. The 9% line is buried in an earnings report. The Bing one-billion-MAU number is a cleaner headline.

And part of it is that the people most affected — small and mid-sized publishers — don't have a trade press of their own. Search Engine Journal serves SEO professionals, most of whom work agency-side or in-house at brands. The publisher class that's actually getting squeezed has no equivalent voice. So the story doesn't get told.

That's the loop. And we built it.

What this means if you're not a publisher

Most of my readers aren't publishers. You run a service business, an e-commerce operation, a SaaS product. You might think the Google Network decline is somebody else's problem. It isn't, and here's why.

First, the publishers being squeezed are the same publishers that historically reviewed your products, ranked your services in best-of lists, and sent you referral traffic. As that ecosystem thins out, the surfaces where your brand can be discovered, written about, and recommended are also thinning out. The AI citation game everyone is suddenly obsessed with depends on a healthy publisher web — that's where the citations come from. If the publisher web shrinks, the citation pool shrinks, and the assumptions underneath every "build authority through earned media" strategy get harder to execute.

Second, the brands that are winning in Cyrus Shepard's analysis — the ones with proprietary assets, task completion, and direct destination demand — are winning partly because they don't depend on the bleeding intermediary layer. If your business currently depends on Google sending people to a publisher who recommends you, you have a longer chain of dependencies than you should. Each link in that chain is getting weaker.

Third, the advertising substitution effect is real. As open-web inventory becomes less attractive, ad budgets concentrate on Google's owned surfaces and the walled gardens. Which means the auctions you compete in for paid traffic — Google Ads, in particular — are going to keep getting more expensive. You're already feeling this if you run a Google Ads account. The macro story is that there's nowhere else for that money to go.

The honest limits

A few things this piece isn't claiming.

It isn't claiming that Google Network revenue decline directly equals publisher revenue decline. The Network segment is one slice of the publisher revenue picture. Many publishers have moved to header bidding, direct deals, subscriptions, and other monetisation models that don't show up in Alphabet's report. The 9% number is a signal about Google's brokered inventory specifically, not the whole publisher economy.

It also isn't claiming the decline is purely AI-driven. Some of it is — AI Overviews shifting where queries resolve is a real factor. But there are concurrent pressures: privacy changes affecting ad targeting, the slow death of third-party cookies, brand safety concerns, and the general maturation of the programmatic market. The AI story is a contributor, not the only cause.

And it isn't claiming Google is doing anything illegitimate by monetising its own surfaces more efficiently. That's what businesses do. The interesting question is what happens to the ecosystem that grew up around the assumption that Google would keep sending traffic to other people's sites in volume — because that assumption is now visibly wrong, and the implications haven't been priced in by most of the industry.

Where this leaves us

If you build a content-driven business in 2026, you have to take the Google Network decline seriously as a planning input. Not as a panic signal — the numbers don't support panic — but as a structural fact about which way the wind is blowing.

Most strategy decks being shown to UK businesses right now are operating on assumptions the data has just contradicted.

The implications track closely with what Cyrus Shepard's analysis already showed: the businesses that win are the ones that don't depend on Google as a traffic intermediary to a third party. They have their own product, they let users complete the task on-site, they have proprietary assets, they have brand demand that pulls users in directly. Everyone else is trying to extract value from a chain of dependencies that's getting shorter, more expensive, and less reliable every quarter.

The SEO industry will keep selling AI citation tracking and Preferred Sources optimisation, and some of that work is genuinely useful. But the harder, less marketable truth in this earnings report is that the open web's monetisation model — the one that funded twenty years of independent publishing — is in measurable decline, and there's no sign of a successor model that puts the money back into the same hands.

Most strategy decks being shown to UK businesses right now are operating on assumptions the data has just contradicted. The companies that update fastest will be fine. The ones still planning around 2022 economics won't be.