DuckDuckGo reported this month that US app installs jumped an average of 18% week-over-week in the six days following Google IO 2026, as a growing number of users actively sought out search engines that do not force AI-generated results into every query. The figures arrived as a YouTube segment from the WAN Show - a weekly technology discussion hosted by Linus Sebastian and Luke Lafreniere and published on June 2, 2026 - put the trend into plain language, with the hosts citing the DuckDuckGo data and discussing whether paid, ad-free alternatives like Kagi now make more practical sense than staying with Google.

The signal is narrow in absolute terms. DuckDuckGo holds roughly 2% of the US search market, and an 18% weekly rise in installs moves a small base. But the timing and the specificity of the behavioral data make it harder to dismiss as noise. The spike began immediately after Google IO, where the company announced a broad expansion of AI Overviews and a tighter integration with AI Mode. The correlation is direct.

What DuckDuckGo's data shows

According to DuckDuckGo, US app installs averaged 18% growth week-over-week across the six days that followed the Google IO conference. App analytics firm Apptopia independently corroborated a 29% increase in average daily US downloads and a 12% increase globally over the same window.

Visits to noai.duckduckgo.com - DuckDuckGo's dedicated landing page for users who want to search without any AI involvement - climbed an average of 23% week-over-week, with a peak of 27.7% on May 24. That page's existence is itself a product statement: DuckDuckGo has built and promoted an explicit opt-out from AI features, a control mechanism that Google does not offer within standard Search.

DuckDuckGo CEO Gabriel Weinberg addressed the trend directly. "Google is force-feeding AI with no way to opt out," Weinberg said. "As a result, their results are getting worse, not better. We want to be the place that puts users in charge and allows them to decide how much or how little AI they want."

The framing from Weinberg is not new. In September 2025, he criticized the proposed Google antitrust remedies as insufficient, arguing that Google "will still be allowed to continue to use its monopoly to hold back competitors, including in AI search." The install data from May 2026 gives that argument a more concrete behavioral dimension.

The Google IO context

Google IO 2026 ran on May 19 in Mountain View, California. The event introduced Gemini 3.5, a new search box architecture, and agentic capabilities that allow AI to complete multi-step tasks across the web on a user's behalf. AI Mode had, by the time of the conference, surpassed one billion monthly users globally, with query volumes doubling every quarter and average query length tripling relative to traditional keyword search.

The absence of an opt-out has been a consistent pressure point. As PPC Land has documented since May 2024, there is no built-in setting to disable AI Overviews entirely within standard Google Search, though a "Web" filter was introduced as a partial workaround. AI Overviews now appear on approximately 16% of Google search results pages. When they do, the first organic link loses an average of 34.5% of clicks, according to Ahrefs research comparing 300,000 searches from March 2024 and March 2025.

The WAN Show hosts drew a sharper picture of what that experience looks like at the individual level. One host described a search session in which a colleague - scrolling past sponsored results that filled the entire above-the-fold area - reached the bottom of the page only to encounter more ambiguous or commercial listings, then had to scroll back up to find what he originally needed. A separate anecdote involved the AI search function returning an incorrect time zone result - off by one hour - nearly causing a missed broadcast start time. The host noted that the error appeared on a repeated search using the same query, ruling out a typing mistake.

These are not system-level statistics. They are individual failure cases. But they illustrate the gap between Google's aggregate user satisfaction data - which the company has consistently cited in earnings calls and press statements - and the on-the-ground experience that drives users to look elsewhere. PPC Land recently examined that gap in depth, noting that a LinkedIn analysis by Carl Hendy found every public claim about user preference for AI Search traced back to Google's own internal data or a single Google-commissioned Ipsos survey.

Kagi: the paid alternative gaining attention

The WAN Show segment spent considerable time on Kagi, a subscription-based search engine that charges users directly and carries no advertising. The hosts described it as a known but not yet widely adopted option, with one host saying he had discussed it on the show before and been considering switching.

According to Kagi's published pricing, the starter plan costs $5 per month and includes 300 searches and 300 AI interactions per month using standard models. The professional tier sits at $10 per month with unlimited searches. A family plan covers six users for $18 per month when billed annually - $216 per year, or $3 per user per month under that configuration. Kagi also offers 100 free searches for new users, which allows a basic evaluation before committing to a subscription.

Several distinctive features came up in the discussion. Kagi's Small Web is a curated index that, according to the platform, "prioritizes authentic human voices" through "only non-commercial sites submitted by real people." The feature is designed to surface content produced for reasons other than commercial gain - a structural contrast with results that have been search-engine-optimized or sponsored. Kagi also offers an AI philosophy statement explaining how and why it integrates AI features, and users can choose not to engage with those features at all.

The business model is the practical distinction. Because Kagi charges users rather than advertisers, it has no structural incentive to rank paid results prominently or to maximize time-on-platform. One commenter on the WAN Show video noted that Kagi does not bill users who do not use their plan in a given month. The engine reportedly sources its index primarily through Bing and Brave APIs, supplemented by its own listings - a common approach for smaller entrants that lack the infrastructure to crawl the web independently.

For comparison, Google maintains 73.7% of US desktop search market share as of Q4 2025, with DuckDuckGo at 1.26% in the same dataset. The gap is vast. But the behavioral question the WAN Show raised is a different one: not whether DuckDuckGo or Kagi can displace Google at scale, but whether a meaningful segment of users - those who search frequently, value accuracy, and have some tolerance for paying for digital services - might rationally choose a paid alternative.

The sponsored results problem

A recurring theme in both the WAN Show discussion and the audience comments was the increasing density of sponsored content in Google Search, and the difficulty of distinguishing paid from organic results. One host described Google's sponsored results as "very clearly kind of malicious at this point," pointing to a session where a colleague was unable to navigate past above-the-fold paid listings to reach a relevant organic result.

A commenter on the video described a navigation search for a veterinary clinic in which Google's top result - a sponsored map entry - pointed to a different clinic entirely. The user only avoided navigating to the wrong location because they double-checked before tapping the start button.

These experiences connect to a broader industry pattern. Google AI Overviews reduce organic CTR by 61% and paid traffic by 68% on queries where they appear, according to Seer Interactive analysis of 3,119 search terms across 42 client organizations from June 2024 through September 2025. Pew Research Center data from July 2025 found users clicked on traditional results 47% less frequently when AI summaries were present. At the same time, sponsored formats have expanded into AI Mode itself, with new ad formats announced at Google Marketing Live on May 22, 2026 placing commercial content inside AI-generated answers.

The combined effect - AI-generated summaries absorbing informational clicks, sponsored results occupying visible real estate, and AI Mode ads embedded within conversational responses - creates a page where organic, unsponsored, non-AI content occupies a shrinking share of the visible surface.

DuckDuckGo's position and limitations

DuckDuckGo offers a mixed picture. The company launched an anonymous AI chat service in June 2024 at duck.ai, allowing users to interact with multiple large language models without conversations being linked to their identity. In July 2025, it added a feature letting users filter AI-generated images from search results. It also redesigned its browser interface in July 2025 to streamline privacy controls. The company does offer its own AI features - it is not anti-AI in principle - but positions user choice as the governing design value.

The structural limitation is one the hosts acknowledged. DuckDuckGo does not have the raw data volume of Google. Its search results come from syndicated sources, primarily Bing, rather than a proprietary index at Google's scale. One host noted that when trying to "actively find the right result rather than just force-feed whatever sponsored garbage you're being paid for, you can overcome a lot of that" - describing a demonstration by a colleague that impressed him despite DuckDuckGo's smaller data footprint.

There is also a contractual wrinkle worth noting. DuckDuckGo's browser architecture blocks third-party trackers, but maintains contractual restrictions regarding Microsoft services, including Bing and LinkedIn Ads, as a condition of its search syndication agreement. This means tracker blocking on non-Microsoft domains is complete, while Microsoft-related services receive different treatment. For users specifically concerned about Microsoft's advertising ecosystem, this is a relevant limitation.

DuckDuckGo peaked at 1.93% US market share in January 2026 before falling back to 1.21% by March, according to Datos data. The install spikes after Google IO represent a shift from a small and volatile base.

What the advertising industry should note

For marketers and advertisers, the user behavior documented this week carries implications beyond competitive search dynamics. The WAN Show hosts noted that a growing segment of technically informed users are willing to pay to escape ads entirely - not just to use an ad-free tier of a platform they already use, but to switch to a fundamentally different product category. Kagi's model, if it attracts sufficient volume, is structurally invisible to advertisers. There is no inventory to buy, no audience to target, and no measurement to run.

The SISTRIX data published via PPC Land in May 2026 showed that click-through rates at position one in Google Search have fallen from 27% to 11% as AI features expand. Queries in AI Mode have more than doubled every quarter. Average query length has tripled. The advertising formats that exist today - keyword-triggered, position-based, text-linked - were designed for a different kind of search than the one Google is now building. Whether users leave Google entirely or simply engage with it differently, the question of how to reach them is becoming more complex.

The WAN Show segment, with 673,000 subscribers and published June 2, 2026, is available at this link. It is not a primary data source - it is a public discussion that surfaced and contextualized data DuckDuckGo itself had published. But as a document of how technically engaged users are currently thinking about Google Search, it reflects a sentiment the install numbers give quantitative form.

Timeline

  • May 14, 2024 - Google launches AI Overviews in the United States at Google IO 2024, placing AI-generated summaries at the top of search results pages with no opt-out available in standard Search.
  • June 9, 2024 - DuckDuckGo launches anonymous AI chat at duck.ai, allowing users to interact with AI models without conversations being tied to their identity.
  • August 2024 - US District Judge Amit Mehta rules that Google illegally maintained monopolies in general search and search text advertising markets.
  • April 21, 2025 - Cloudflare's Q1 2025 search report confirms Google at 87% global search market share; DuckDuckGo at 1.501%.
  • April 2025 - A Virginia court rules Google violated antitrust law by monopolizing digital advertising technology markets.
  • July 24, 2025 - DuckDuckGo redesigns its browser interface with streamlined privacy controls; contractual restrictions on Microsoft tracker blocking remain in place.
  • July 25, 2025 - DuckDuckGo adds a feature allowing users to filter AI-generated images from search results.
  • September 7, 2025 - DuckDuckGo CEO Gabriel Weinberg criticizes proposed Google antitrust remedies as failing to address AI search competition.
  • October 10, 2025 - UK CMA designates Google with Strategic Market Status, covering Google Search including AI Overviews and AI Mode.
  • November 5, 2025 - Seer Interactive analysis published via PPC Land finds AI Overviews reduce organic CTR by 61% and paid traffic CTR by 68% on affected queries.
  • January 2026 - DuckDuckGo peaks at 1.93% US search share before falling to 1.21% by March, according to Datos.
  • March 4, 2026 - PPC Land reports Q4 2025 desktop search data: Google at 73.7% US market share, DuckDuckGo at 1.26%.
  • May 17, 2026 - SISTRIX data shows position-one CTR in Google Search has fallen from 27% to 11% as AI features expand, as covered by PPC Land.
  • May 19, 2026 - Google IO 2026 announces Gemini 3.5, agentic search features, and broader AI Mode rollout; AI Mode surpasses one billion monthly users globally.
  • May 20-25, 2026 - DuckDuckGo records an average 18% week-over-week rise in US app installs; noai.duckduckgo.com traffic rises 23% on average, peaking at 27.7% on May 24.
  • May 22, 2026 - Google Marketing Live announces new AI search ad formats including conversational discovery ads and highlighted answer ads embedded within AI Mode responses.
  • May 26, 2026 - Apptopia independently reports a 29% increase in average daily US DuckDuckGo downloads and a 12% global increase over the same post-IO window.
  • May 28, 2026 - PPC Land examines the evidentiary basis for Google's claim that users prefer AI Search, finding all public claims trace to Google-owned or Google-commissioned data sources.
  • June 2, 2026 - WAN Show segment published on YouTube discusses DuckDuckGo install data and evaluates Kagi as a paid alternative, reaching 673,000 subscribers.

Summary

Who: DuckDuckGo, its CEO Gabriel Weinberg, the WAN Show hosts Linus Sebastian and Luke Lafreniere, and the broader population of users who installed DuckDuckGo or visited its AI-free search page in the week following Google IO 2026.

What: DuckDuckGo reported an 18% average week-over-week increase in US app installs across six days following Google IO 2026. Visits to noai.duckduckgo.com rose 23% on average. A YouTube discussion segment published today surfaced and contextualized that data, and also examined Kagi - a paid, ad-free search engine starting at $5 per month - as an alternative that gives users explicit control over AI involvement in their searches.

When: The install spike occurred between approximately May 20 and May 25, 2026, immediately following Google IO 2026 on May 19. The WAN Show segment discussing the data was published June 2, 2026.

Where: The install data covers the United States. DuckDuckGo's AI-free landing page is accessible at noai.duckduckgo.com. The WAN Show discussion is published on YouTube at https://www.youtube.com/watch?v=xet2ZyZKAb4. Kagi is available at kagi.com.

Why: Google IO 2026 expanded AI Overviews and deepened AI Mode integration without introducing an opt-out mechanism for AI features in standard Search. A segment of users responded by installing an alternative with an explicit AI-free option. The WAN Show discussion reflects wider frustration with the density of sponsored content in Google Search results and documented accuracy failures in AI-generated answers, and examines whether a paid subscription model - one with no advertising revenue to protect - is now a rational choice for frequent searchers.