How much traffic comes from AI search engines vs Google?

As of early 2026, Google still dominates raw search query volume — roughly 8.5 billion queries per day across Search globally, per company disclosures and Similarweb estimates. AI search engines combined (ChatGPT, Claude, Perplexity, Copilot, Meta AI, Gemini standalone) handle an estimated 1–2 billion queries per day, with Perplexity alone at ~500 million per month and ChatGPT's search-enabled queries running into the high hundreds of millions per month.

Referral traffic tells a different story. AI engines drive far fewer click-through visits than their query volume suggests, because most queries terminate inside the answer. Referral-traffic data from Similarweb and Cloudflare Radar (Q4 2025) shows AI search engines drove roughly 1–3% of referral traffic to most consumer publishing sites, vs. 30–60% from Google Search. But the trajectory is the load-bearing fact: AI referral traffic grew 4–6x year-over-year, while Google referral traffic was flat or declining for the same period.

The right way to think about it: AI search is a brand and consideration channel, not a traffic channel — yet. You measure success in citation share and brand-query lift, not in GA4 sessions. The traffic-driving share will grow as AI engines add product links, citation hovers, and shopping integrations, but the leading indicator today is whether you're present in the answer, not whether the user clicked through.