DuckDuckGo published install and visit data this week showing an 18.1% average week-over-week increase in US app installs during May 20 to May 25, a period that overlapped directly with Google I/O and its high-profile AI Mode push. The peak came on May 25, when daily installs ran 30.5% above the prior week’s baseline. The data, reported by TechCrunch on May 26, gives the first measurable shape to a user reaction that Search Shift flagged on May 21 as a directional trend based on app-store ranking movement. The numbers are now confirmed and specific.
The iOS figures deserve separate attention. iPhone users averaged a 33% week-over-week install increase across the same window, with a single-day peak of 69.9% on May 25. That kind of one-day spike is not organic word-of-mouth; it typically requires a trigger event visible to a broad audience. Google I/O’s keynote announcement of AI Mode as a default-path experience for US users was that event.
The noai.duckduckgo.com variant, a version of the search engine that routes queries without any AI-generated elements, saw visits grow 22.7% week-over-week on average, peaking at 27.7% on May 24. That detail is the most analytically useful number in the dataset. It signals that the rejection is targeted: users are not fleeing search engines as a category; they are avoiding AI in search results specifically. A user who bookmarks the noai subdomain has made a deliberate product choice, not a passive one.
That distinction separates this moment from two prior rival-engine surges worth benchmarking against. During the Search Plus Your World backlash in 2012, DuckDuckGo saw install spikes driven primarily by privacy concerns about Google’s social-data integration in results; the complaint was about personalization and data collection, not about answer format. During Bing’s ChatGPT integration in early 2023, Bing itself was the AI-forward alternative, picking up users who wanted more generative capability, not less. The May 2026 surge runs in the opposite direction: users moving away from AI-forward results toward a search experience that explicitly avoids them.
Google’s overall position in search is not at risk from this data. Google holds roughly 90% of global query share, and a week of elevated DuckDuckGo installs does not change that structure. The significance is in the signal, not the scale. iOS users in the US, a high-value demographic for search monetization, are showing a measurable preference response within days of a Google product announcement. That is a faster feedback loop than the search industry typically sees from user-behavior data.
The long-tail implication for share-of-search measurement is practical. Brands and publishers that track branded query volume in Google Search Console may not see any change; DuckDuckGo’s install growth happens in a separate measurement silo. The more useful indicator to watch is branded direct traffic and any increase in referral traffic tagged to duckduckgo.com. If the install surge converts to sustained query volume on DuckDuckGo, some portion of that traffic will eventually appear in site analytics as a new referrer, not as a shift in Google-sourced clicks.
DuckDuckGo’s own attribution, per TechCrunch’s reporting, points directly to Google I/O’s AI Mode announcements as the driver. The company’s stated growth in the noai variant reinforces that framing. What the data does not answer is whether these installs represent permanent switching or trial behavior; a user who installs an alternative engine after a headline-grabbing product event often returns to the default within two to four weeks.
SEO teams with significant iOS traffic should establish a May 2026 baseline for DuckDuckGo referral volume now, before a sustained shift is visible in aggregate data, so they have a comparison point if the pattern holds through June.
Reported by TechCrunch on 2026-05-26, written by Rebecca Bellan.
