Less than a week after deploying the compute-heavy Gemini 3 model to power its new “Thinking Mode,” Google has moved to monetize the expensive reasoning process by inserting advertisements directly into the interface.
Confirmed by the company following user reports on Thursday, “Sponsored” placements now appear at the bottom of complex reasoning chains for some users, mimicking organic results.
This rapid shift signals the end of the subsidy period for agentic search, as the tech giant attempts to offset the high inference costs of its “Answer Engine” while facing fresh advertiser backlash over its automated campaigns.
From ‘Labs’ to Reality: The Mechanics of Thinking Mode Ads
Transitioning from the experimental “beaker” icon to standard search results, Google has formalized the presence of ads with tests for its AI-powered “Thinking Mode.”
Following the November 18 Gemini 3 integration, the test-deployment suggests the architecture was designed with immediate monetization in mind.
Visually, the advertisements are currently sequestered at the bottom of the reasoning chain rather than interrupting the generated text. As documented by SEO Consultant Brodie Clark:
“Similar to the ads that I had originally spotted showing within AI overviews earlier this year, the placement is at the bottom of the response. And as you can see, the ad results are labelled with the ‘sponsored’ tag, appearing in a similar way to the organic link cards directly above.”
Each unit is clearly labeled with a “Sponsored” tag, mimicking the design language of organic link cards to reduce visual friction.
Clark was among the first to document the shift, noting the permanence of the implementation. He observed that “the time has come. Google is now showing ads within AI Mode for the first time ever.”
The time has come. Google is now showing ads within AI Mode for the first time ever.
Similar to the ads that I had originally spotted showing within AI overviews earlier this year, the placement is at the bottom of the response.
And as you can see, the ad results are labelled… pic.twitter.com/JZa7g8rS2N
— Brodie Clark (@brodieseo) November 21, 2025
A Google spokesperson confirmed in a statement to Android Authority that this is an expansion of tests announced in May, though the visibility has spiked significantly in the last 48 hours. Strategically, the placement appears designed to minimize immediate user rejection while establishing the real estate for future inventory.
The Economic Imperative: Funding the ‘Reasoning’ Engine
Shifting to “Thinking Mode” introduces a significant spike in inference costs, as Gemini 3 performs multi-step reasoning rather than simple keyword retrieval.
Industry analysts view the rapid ad rollout as a necessary mechanism to subsidize these compute-heavy queries, ending the “free lunch” phase of agentic search.
To justify the likely lower click volume, platforms are pivoting to a “quality over quantity” narrative. Microsoft has released data regarding claims of higher conversion rates, arguing that AI-driven referrals perform significantly better than traditional search traffic. The company’s analysis states:
“Even though AI referrals still represent under 1% of total visits, these users consistently display higher intent and engagement, signing up, subscribing, or purchasing at far greater rates.”
Google executives have laid the groundwork for this argument, framing AI interactions as high-intent “discovery” moments.
Dan Taylor, VP of Global Ads, previously described the format as an “expansive opportunity to introduce advertisers—and put them in front of consumers in places where they’re open to discovering new things.”
However, this “value” argument clashes with the reality of publishers who rely on volume for ad impressions.
Advertiser Friction: The ‘AI Max’ Shadow
Introducing new inventory arrives at a precarious moment for Google’s relationship with buyers, who are currently revolting against automated “AI Max” campaigns.
Advertisers have reported severe inefficiencies with Google’s latest AI automation tools, citing sky-high costs without commensurate performance.
Independent reports indicate Cost Per Acquisition (CPA) for AI Max campaigns has surged by nearly double in some verticals. Xavier Mantica, an advertiser documenting the trend, noted that “AI Max campaigns are showing 90% higher cost per conversion than phrase match.”
Eroding trust, the “black box” nature of these tools has made marketers wary of automatic placements in new surfaces like “Thinking Mode.” While Google cites internal case studies like L’Oréal showing conversion lifts from the AI Max for Search campaigns launch, third-party data suggests these may be outliers.
Marketers fear that “Thinking Mode” slots will become another opaque inventory source where budgets are drained with little transparency. If the “Sponsored” cards at the bottom of reasoning chains fail to deliver the promised “high intent” conversions, the economic model of agentic search could falter.
Strategic Divergence: The Agentic Arms Race
Google’s aggressive monetization highlights the considerable structural advantage of its existing ad stack compared to emerging rivals. Brodie Clark emphasized this disparity, noting that “this marks a significant difference between Google’s AI Mode and the likes of ChatGPT, with ChatGPT not having even close to the same infrastructure for ads.”
Competitor Perplexity has taken a divergent path, having paused ads in October 2025 to “rethink” its strategy after user pushback. Instead of direct injection, Perplexity is pursuing partnerships, evidenced by its recent $400 million deal to power search inside Snapchat.
OpenAI remains in a tentative position, lacking the mature ad server infrastructure to monetize “Thinking” models immediately. Google is effectively betting that its dominance allows it to force ads into the agentic experience before competitors can offer a viable, ad-free alternative.
Risks remain that if the ad load increases—moving from the bottom to the middle of the reasoning chain—users may defect to cleaner interfaces.
