Within days of Google unveiling its most aggressive push yet to replace traditional search results with AI-generated answers, a measurable number of Americans went looking for the exit. DuckDuckGo, the privacy-focused search engine that keeps AI features off by default, recorded an average 18.1% week-over-week increase in U.S. app installs between May 20 and May 25, 2025, according to data the company shared with multiple newsrooms. On May 25 alone, installs spiked 30.5% compared with the same day the prior week.
The timing is hard to dismiss as coincidence. Google’s annual I/O developer conference ran May 20 and 21, and the centerpiece announcement was a sweeping expansion of AI-generated summaries across Google Search. For millions of users who have spent years scanning ranked blue links and deciding for themselves which to click, the redesign felt less like an upgrade and more like a product they never asked for.
The numbers and where they come from
DuckDuckGo’s figures are self-reported. The company briefed several outlets with the same dataset, and the topline numbers align across Yahoo’s coverage, Indian Express reporting, and a PCMag analysis: an 18.1% average week-over-week increase across both iOS and Android, sustained over a six-day window, with the sharpest acceleration at the end of the period. No third-party analytics firm such as Sensor Tower or data.ai has publicly corroborated the figures, so they should be treated as indicative rather than independently verified.
A second metric from the same briefing adds depth. DuckDuckGo’s browser-based search page, where users go directly to the site without downloading an app, saw 22.7% week-over-week growth in traffic during the same window. The gap between the two numbers suggests a familiar adoption pattern: some users tested DuckDuckGo in a browser tab before committing to an app install, the digital equivalent of browsing before buying.
What DuckDuckGo has not shared is the raw install count behind those percentages. That matters. An 18% jump from 50,000 weekly installs is a different story than an 18% jump from 500,000. DuckDuckGo holds roughly 2.5% to 3% of the U.S. search market, according to StatCounter data from early 2025, which means even a strong growth week leaves Google’s roughly 89% share functionally untouched. The percentage gains are real; their competitive significance is a separate question.
Why this backlash feels different, and why it might not be
Google has weathered user frustration before. When the company first rolled out AI Overviews in May 2024, social media lit up with complaints about inaccurate AI-generated answers appearing above organic results, and DuckDuckGo reported a similar, smaller bump in interest. That wave subsided within weeks as Google refined the feature and users adapted or simply scrolled past the summaries.
This time, the scale of Google’s commitment is larger. At I/O 2025, the company signaled that AI-generated answers would become the default experience across its core search product, not just an optional layer. For users who viewed AI Overviews as an annoyance they could ignore, the prospect of AI becoming the primary interface raised the stakes. The install data suggests that at least some users responded by acting rather than just complaining.
Still, online frustration does not always translate into lasting behavior change. Vocal critics on platforms like X and Reddit may not represent the hundreds of millions of people who use Google without strong feelings about how results are formatted. Switching a default search engine requires deliberate effort, and most people do not bother. The one-week window covered by DuckDuckGo’s data cannot reveal whether these new users stayed or drifted back to Google within days.
The competitive landscape beyond DuckDuckGo
DuckDuckGo is not the only alternative benefiting from dissatisfaction with AI-saturated search. Brave Search, Startpage, and the subscription-based Kagi have all positioned themselves as options for users who want traditional link-based results or more control over AI features. None of these competitors have published post-I/O install data comparable to DuckDuckGo’s, but the broader pattern of niche search engines marketing themselves as refuges from AI is unmistakable.
DuckDuckGo’s particular advantage is brand recognition. It is the most widely known Google alternative in the U.S. and is already integrated as a default option in browsers like Safari and Firefox. That built-in visibility means it is often the first place frustrated users land, whether or not they ultimately stay. The company has leaned into this positioning by emphasizing that its search results remain link-first and that any AI-powered features require users to opt in, a direct contrast with Google’s approach of making AI the default.
Google, for its part, has not publicly responded to DuckDuckGo’s claims or released any counter-data about its own retention trends. The company processes billions of queries daily, and a strong week for a competitor with low single-digit market share may not register as material movement internally. But the pattern of backlash-driven experimentation is worth watching, particularly if Google continues expanding AI features throughout the rest of 2025.
What the data can and cannot tell us
The strongest takeaway from DuckDuckGo’s numbers is that a measurable group of U.S. users actively sought alternatives to Google’s AI-first search experience in the days immediately following I/O 2025. Two metrics, app installs and browser traffic, moved in the same direction after the same external event, which strengthens the case that something real happened even if the precise cause cannot be isolated with certainty.
What the data cannot tell us is whether this is the beginning of a sustained shift or a brief flare of experimentation. History favors the latter. Google’s dominance rests on deep integration with Android, Chrome, and default search agreements worth billions of dollars annually. Displacing that kind of structural advantage requires more than a week of elevated app downloads.
For now, the numbers are best read as a focused snapshot: a specific group of users, in a specific week, reacting to a specific product change. Whether that snapshot becomes a trend depends on what Google does next with AI in search, and on whether alternatives like DuckDuckGo can convert curious downloaders into daily users. As of late June 2025, that question remains open.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
