Every Major AI Assistant Recommends Its Parent Company's Models More Often Than Other Engines Do -- Except One. New 5W Research Finds.

5W’s AI Companies AI Visibility Index 2026 (32,200 prompts, Jan–May 2026) finds self-citation is broadly present: ChatGPT recommended OpenAI models 2.0x more often than other engines, while Claude is the outlier recommending Anthropic models only 1.2x more often (lowest lift tested). Citation share is led by OpenAI at 24.6% vs Anthropic 14.8%, Google DeepMind 13.7%, Meta AI 9.7%, and xAI 5.7%, with the top five totaling 68.5% of cited responses. The index also notes Anthropic over-indexes in safety queries (31.2% vs 14.8% overall), and flags an AI communications asymmetry centered on GitHub/ArXiv as primary “source material.”

Analysis

The investable takeaway is not the self-citation metric itself; it is that AI discovery is starting to behave like a distribution moat, and the winner is whoever controls the default routing layer. That structurally favors GOOGL more than META because Google can convert visibility into retained search queries, higher ad load, and better enterprise mindshare across search, browser, and model surfaces. The economic impact is medium-term: this matters only if the visibility edge shows up in query share, product adoption, or cloud attach rates over the next 1-3 quarters.

For META, the signal is weaker but still relevant as a relative positioning issue. Under-indexing in AI discovery does not break the ad business, but it suggests Meta’s AI spend is still more of a defensive capex race than a compounding platform advantage, which can cap multiple expansion if investors start demanding proof of monetization efficiency. The more durable second-order beneficiaries are vertical AI application layers, because domain-specific prompts appear to convert into higher citation share; that is a positive for private names, not enough yet to underwrite public proxies.

The contrarian view is that this is mostly branded-SEO theater, not revenue. The methodology is survey-based and can overstate moat versus actual user behavior; the key falsifier is whether traffic, conversion, or enterprise win rates move in the next earnings cycle. Regulatory risk is the cleaner 1-3 month catalyst: if self-preferencing scrutiny rises, GOOGL’s perceived advantage could flip into a headline overhang faster than it converts into fundamentals.

Key Decisions for Investors

• No outright trade on the headline alone; wait for earnings evidence on query share / ad RPM before assigning value to GOOGL's AI visibility edge.
• If you need a relative-value expression, buy GOOGL / sell META on a 3-5% spread widening in the next 1-3 weeks; thesis is ecosystem leverage, not a near-term EPS beat. Falsifier: no improvement in Search/Cloud commentary or rising antitrust noise.
• Use this as a watch item on ARBU and INSO only if they are liquid AI-discovery / GEO proxies; do not buy without evidence of ARR or customer adoption, because the article is sentiment-led rather than earnings-led.
• Set an alert for any regulatory commentary on AI self-preferencing in search/overviews over the next 30-60 days; that is the cleanest catalyst to short GOOGL on weakness if scrutiny escalates.
• For META holders, trim into strength if management continues to emphasize AI spend without a visible monetization bridge over the next 1-2 quarters; the risk/reward improves only when AI starts reducing CAC or lifting engagement measurably.