DuckDuckGo launched in 2008 with a pitch that sounded almost quaint at the time: a search engine that wouldn’t track you. For more than a decade, that positioning earned the company a loyal niche, a steady drumbeat of press coverage during every Google privacy scandal, and almost no movement on the only metric that mattered. Market share stayed stuck in the low single digits while Google’s grip on U.S. search held firm above 90%.
Then Google unveiled its AI-first Search overhaul at I/O, and something shifted. DuckDuckGo recorded a 30% surge in installs: six consecutive days of double-digit growth, including across a Memorial Day weekend when the company typically sees a traffic dip. According to TechCrunch, U.S. app installs jumped roughly 30% on a week-over-week average in late May, with iOS leading the spike.
That number is the story. Not because it threatens Google’s 90% share, but because it puts a concrete figure on something nobody has previously been able to measure: the size of the captive market inside a search monopoly that would actively prefer not to be there.
The trigger: Google replaces blue links with an AI agent
At its annual developer conference, Google announced changes to its search experience, moving away from traditional lists of blue links toward an AI agent that answers queries, executes tasks, and runs background monitoring agents. The redesign folds AI Overviews into the default experience, and critically, removes the user-facing option to opt out.
Critics have argued the redesign could kill the open web by removing the outbound clicks that fund publishers. But the structural complaint is narrower than the ideological one: users no longer have the option to receive a plain list of results.
What a 30% surge on a 2% base actually measures
DuckDuckGo holds roughly 2% of the U.S. search market. A 30% install surge on a 2% base is not an inflection point in market share, and that is precisely what makes it analytically useful. It is a clean measurement of the latent anti-default market: the share of users who, when an opt-out disappears, are willing to take the friction of installing a competing app rather than accept the new experience.
Visits to noai.duckduckgo.com, the AI-free search page that disables AI-assisted answers and AI-generated images by default, showed parallel growth.
The signal is consistent: this is not a generalised AI backlash but a specific reaction to the removal of choice.
Why the size of this market was previously invisible
The defection moment is taking place inside a regulatory environment that has spent two years scrutinising exactly how Google maintains its dominance. During Google’s 2023 search antitrust trial, DuckDuckGo CEO Gabriel Weinberg testified that Google’s exclusive default search contracts harmed DuckDuckGo’s ability to pitch itself as the default on other browsers. That testimony is the key to reading the current numbers. For years, the standard defence of Google’s market share has been that users could leave if they wanted to. They simply don’t. The 30% install spike complicates that argument. The dissatisfied population was always there; it was suppressed by distribution defaults, not by genuine preference. When the product changed sharply enough to overcome the friction of switching, a measurable slice moved within days. Silicon Canals has previously reported on €25B in damage claims filed against Google in UK and Dutch courts over adtech practices, part of the same broader question: how much unilateral control should a default provider hold when its product decisions affect the majority of the world’s online queries with no native opt-out?
What this episode reveals — and what it doesn’t
The 30% number describes the size of the addressable opt-out market in a single week, under a single trigger. It does not describe a shift in the underlying distribution machinery that has kept Google at the top. Sustained share gains in search have historically required either a default-distribution win or a regulatory intervention, not user preference alone.
But the install data should change the tenor of every pending antitrust case against Google. The default-contracts argument has always rested on a counterfactual that critics could only assert: that absent the contracts, a meaningful number of users would choose differently. That counterfactual is no longer hypothetical. When Google removed a single opt-out, the latent demand for an alternative surfaced inside a week, at a scale that dwarfs DuckDuckGo’s existing footprint.
The implication is uncomfortable for Google and clarifying for regulators. Most of the 90% share is structure, not preference. The company has spent years arguing the two are indistinguishable. They aren’t, and now there is a number to prove it.
