Small publishers lost 60% of search referrals. Google beat a publishers antitrust suit. For brands, the game is becoming a source inside the answer layer.
Google.com now accounts for 17% of all AI Mode citations, up 3x from mid-2025. A new study reveals how Google AI is increasingly citing its own properties over independent publishers.
A new study from SE Ranking analyzing more than 1.3 million citations from Google's AI Mode reveals a trend that should concern every independent publisher: Google is increasingly citing itself.
Google.com now accounts for roughly 17% of all citations in AI Mode responses—approximately three times what it was in June 2025. YouTube ranks as the second most-cited domain overall. Combined, Google's own properties—Search, YouTube, Maps, Business Profiles, and others—make up about one-fifth of all sources cited by AI Mode.
In certain verticals like travel and entertainment, the numbers are even more stark: more than half of citations point back to Google's own pages rather than external publishers.
What the Data Shows
SE Ranking's study examined 68,313 keywords across 20 industry niches, collecting 1,321,398 total citations from AI Mode responses during February 2026. The findings paint a clear picture of consolidation.
The self-citation rate of 17.42% does not sound dramatic in isolation. But context matters. Less than a year ago, that figure sat around 5–6%. The trajectory is steep, and it is accelerating as Google expands AI Mode—which rolled out to all U.S. users at Google I/O 2025 without requiring any signup.
The domain breakdown tells the story:
Google.com: 17.42% of all citations (up from ~5–6% in mid-2025)
YouTube.com: Second most-cited domain overall
Google properties combined: ~20% of all cited sources
Travel and entertainment verticals: Over 50% of citations point to Google pages
As Wired reported, this pattern creates self-referential loops that keep users inside Google's ecosystem while squeezing out the independent publishers whose content trained the underlying models.
The Self-Preferencing Problem
This is not a new accusation against Google. But AI Mode adds a new dimension to it.
In traditional search, Google's results page linked out to external websites. Users clicked through. Publishers got traffic. The exchange was imperfect but functional. With AI Mode, Google generates a synthesized answer directly in the search interface—and when it cites sources, it increasingly cites itself.
The mechanism is straightforward. Google controls both the content layer (YouTube, Maps, Business Profiles, Shopping) and the distribution layer (Search, AI Mode). When the AI retrieval pipeline selects sources, Google's own properties benefit from privileged integration with the Knowledge Graph and structured data systems that power AI Overviews.
Regulators have noticed. The European Commission opened a formal antitrust investigation in December 2025 into whether Google used web publisher and YouTube content for AI purposes without proper compensation—and whether publishers can refuse that use without losing access to Google Search entirely. In the U.S., the Department of Justice secured remedies after Judge Mehta's ruling that Google illegally monopolized search, with provisions explicitly extending to generative AI products like Gemini and AI Mode.
The Publisher Traffic Collapse
The self-citation trend does not exist in isolation. It coincides with a broader collapse in publisher traffic from Google.
According to data from Chartbeat tracking 2,500+ publisher websites, Google search referrals to publishers fell 33% globally—and 38% in the U.S.—year-over-year through November 2025. Publishers surveyed by the Reuters Institute expect traffic to decline an additional 43% over the next three years.
The math is brutal. AI Mode generates answers that reduce click-through to external sites. When users do click a cited source, that source is increasingly a Google property. External publishers lose on both counts.
What This Means for Search Strategy
For brands and publishers competing for visibility in AI-generated answers, the implications are direct:
Google-owned channels matter more. YouTube optimization, Google Business Profile completeness, and Google Merchant Center feeds now influence AI Mode citations. Treat these as first-class search assets.
Entity authority is critical. Content that is clearly grounded in identifiable entities—with structured data, consistent naming, and unambiguous claims—has a better chance of being cited over a Google property that covers the same topic generically.
Diversify discovery channels. Relying exclusively on Google for traffic was already risky. With AI Mode accelerating the shift, building presence across ChatGPT, Perplexity, Claude, and other AI search surfaces is no longer optional.
Track AI citations, not just rankings. Traditional rank tracking does not capture whether your content appears in AI-generated answers. Tools that monitor AI citation share across queries are becoming essential.
The Bigger Picture
Google building products that cite Google is not surprising. It is incentive-aligned behavior from a company that controls the dominant search platform, the dominant video platform, and the dominant mapping platform. The question is whether this behavior constitutes self-preferencing that harms competition—and regulators on two continents are actively investigating that question.
For the rest of us, the strategic takeaway is clear. The era of Google as a neutral traffic distributor is ending. AI Mode is accelerating that transition. Publishers who adapt their content strategy to this reality—optimizing for AI retrieval across multiple platforms, not just Google's blue links—will fare better than those waiting for the old model to return.
It is not coming back.
James Calder is the editor of The Search Signal, covering AI-powered search, generative engine optimization, and the future of brand discovery.
Small publishers lost 60% of search referrals. Google beat a publishers antitrust suit. For brands, the game is becoming a source inside the answer layer.
Perplexity's Personal Computer turns a Mac mini into a persistent, always-on AI agent that works across your local files and apps around the clock — a new category of ambient computing.
