In the first quarter of 2026, Baidu App's monthly active users declined to 655 million, a 10% year-over-year decrease, reflecting the impact of AI search on traditional search.

Article author: Wildcard

Source: GeekPark

In the first quarter of 2026, Baidu App's monthly active users declined to 655 million, a 10% year-over-year decrease.

Baidu is not an ordinary app; over the past 20 years of China’s internet, “Baidu it” has been virtually synonymous with “searching online.” It has been the starting point for traffic, the gateway to information, and the vital water source upon which countless websites depend.

However, in the AI era—users are increasingly inclined to get answers directly from AI assistants rather than clicking through links via search engines.

The whole world shares the same cold and heat.

On the other side of the Pacific, the same script is playing out in an even more dramatic way. A massive "traffic collapse" affecting the entire internet is underway.

And people don't care.

01 No One Survives

In March 2026, the analytics firm Chartbeat provided Axios with a set of exclusive data that sent shockwaves through the entire publishing industry.

This dataset tracks traffic changes over the past two years across thousands of customer websites worldwide using Chartbeat, and the findings are alarming—small publishers (websites with 1,000 to 10,000 daily page views) experienced a 60% drop in search-driven traffic. Medium-sized publishers saw a 47% decline. Large publishers experienced a relatively more modest decline, but still fell by 22%.

The smaller the operation, the harder the fall. The reason is straightforward—large media outlets have brand recognition, apps, email subscriber lists, and direct access to loyal users. Small publishers, by contrast, bet everything on Google search traffic; when that traffic plummets, they have no buffer.

Page views from Google Search decreased by 34% between December 2024 and December 2025. Google Discover also declined by 15% during the same period. The entire Google "traffic faucet" is being tightened.

But for individual websites, the numbers are far worse than these averages.

In a report released by Loopex Digital in May 2026, a list of the most affected websites was published. Business Insider’s traffic plummeted by 48.4%. WebMD’s monthly unique visitors dropped from 122 million to 69.5 million. Stack Overflow, the world’s largest programming Q&A community, saw a 35.5% decline in traffic, while Google search volume fell by 32.4%. Quora declined by 28.1%. Even Unsplash, a stock photo website, was not spared, with monthly users dropping from 26.8 million to 17.7 million—a 34% decrease.

If you look across this list horizontally, you'll notice a disturbing pattern—the most severely cut are precisely the core components of the "internet information infrastructure" over the past decade.

News media (Business Insider, CNN), Q&A platforms (Quora, Stack Overflow), health information websites (WebMD), and image stock platforms (Unsplash)—these websites share a common trait: their fundamental value lies in answering users' questions.

And now, someone else is answering for them.

02The Zero-Click Era

To understand what’s happening, you first need to grasp the concept of “zero-click search.”

This term refers to users entering a question into a search engine and receiving an answer directly on the search results page, without needing to click any links or visit any websites. Approximately 60% of searches worldwide now end in zero clicks. On mobile devices, this figure rises to 77%.

Zero-click searches are not a new phenomenon of the AI era. Google has long been doing the same thing through features like “Featured Snippets,” “Knowledge Panels,” and “People Also Ask”—displaying answers directly on the search results page to reduce the reasons for users to click away from Google. But AI Overviews takes this to an entirely different level.

According to data from multiple studies, Google AI Overviews currently cover approximately 48% of search queries, a 58% year-over-year increase. When an AI Overview appears on a search results page, the click-through rate for the top organic result drops by 34% to 58%—depending on the study—but the trend is entirely consistent.

Seer Interactive's analysis is more aggressive, suggesting that overall click-through rates for natural search dropped by 61% in queries with AI Overview.

In short: your website is still in the search results, its ranking may be unchanged, but your click-through rate may have already halved. This is the most confusing part for many website operators—“I didn’t do anything wrong, but my traffic has plummeted.”

But Google AI Overviews are only half the story.

The other half is an entire emerging "AI search" ecosystem growing rapidly.

ChatGPT now processes over 1 billion search queries per week. Perplexity’s monthly queries have also surpassed 1 billion. Claude, Gemini, and Copilot are all growing rapidly—Goodie’s 2026 AI Search Traffic Report shows that Claude’s share of AI-driven traffic jumped from 11.8% to 18.5% in early 2026, the fastest growth among all AI search engines.

The way these AI products operate differs fundamentally from traditional search engines. Google’s former business model was to “send users to websites”—you search for a question, Google gives you a list of links, you click through, and Google’s job is done. This is a “traffic referral” model.

The logic of AI chat products is to “give you the answer directly.” It reads your content, understands your knowledge, synthesizes a response, and delivers it straight to the user. The user is satisfied and leaves—throughout the entire process, they never need to know where the answer came from or visit any website.

That’s why focusing solely on Google’s traffic changes isn’t enough. What’s truly happening is much deeper than just “Google algorithm updates.”

The entire internet is transitioning from a “network of links” to a “network of answers.”

Over the past 20 years, the fundamental chain of internet information flow has been: creators produce content → search engines index and rank it → users click links → visit websites → content creators receive traffic and advertising revenue. This is a closed loop, with each participant fulfilling a distinct role and having their own interests.

Now, this chain is being broken. The new chain is: creators' content is scraped and learned by AI → AI directly answers users' questions → users leave satisfied → content creators receive nothing.

The central "click" mechanism, which for twenty years has served as the core infrastructure supporting the entire digital publishing, digital advertising, and content marketing industries, is being systematically eliminated.

03 China's "Parallel Universe"

Some might think this is primarily a story about Google and the English-language internet, with little relevance to China.

On the contrary, China is experiencing the same thing, just with different companies taking the traffic.

In the United States, the main force diverting traffic is Google’s own AI Overviews—Google places an AI-generated summary at the top of its search results, keeping users on the search results page instead of directing them to external websites. This is essentially Google diverting its own traffic.

The situation in China is even more intense. It’s not traditional search engines themselves that are being cut off, but an entire wave of independent AI-native apps—Doubao, Qwen, DeepSeek—that are directly stealing users away from traditional search.

The numbers speak for themselves.

According to QuestMobile data, in April 2026, the average monthly usage frequency of search engine apps decreased by 18.8% year-over-year, and average usage duration dropped by 11.8%. During the same period, the number of monthly active users of AI-native apps approached 500 million, with an average of 91 uses per user per month and an average usage duration of 180 minutes. By June, Doubao’s monthly active users had surpassed 528 million.

5.28 billion—this number is approaching Baidu App’s 6.55 billion monthly active users. And the former is still growing, while the latter is still declining.

In its May 2026 survey, QuestMobile found that among the 25 industry apps observed, 28% of industries experienced year-over-year declines in both user frequency and duration, while another 40% saw a decline in at least one metric. Search, automotive news, and travel were the first three industries to feel the impact of AI-driven information distribution. Among online travel users, 69.4% are also using AI-native apps; among automotive news users, this figure is 51.1%.

There’s a particularly striking piece of data here. After being cited by Doubao, Qwen, and DeepSeek, the content from Pacific Auto reached a combined audience of 12.37 million users. That sounds like good news—until you realize this number is 113.8 times the size of Pacific Auto’s own app traffic.

113.8 times. That means over a hundred times more users “consumed” content from Pacific Auto via the AI platform than ever visited Pacific Auto’s website. They asked a question about buying a car in the Doubao chat window, Doubao referenced Pacific Auto’s content to provide an answer, and the user closed the conversation satisfied.

PacAuto contributed knowledge, the AI platform gained users, and PacAuto gained nothing.

The companies taking away traffic may differ, but the underlying logic of erosion is exactly the same—whether it’s Google AI Overviews or Doubao, they’re essentially doing the same thing: consuming the knowledge of content creators, synthesizing answers, and delivering them directly to users while bypassing the original sources.

04 Traffic that never comes back

At this point, some might think: Yes, AI has taken away search engine traffic, but doesn’t AI also cite sources and include links? Aren’t these AI products also driving traffic to websites?

Theoretically, yes. In practice, it’s negligible.

Chartbeat data shows that traffic from AI chat tools increased by more than 200% from December 2024 to December 2025. However, even after this growth, it still accounted for less than 1% of publishers’ total page views.

Contentsquare’s 2026 Digital Experience Benchmark Report provides more precise figures—AI-driven traffic surged by 632% year-over-year, yet still accounted for only 0.2% of all traffic. Moreover, the quality of this 0.2% is concerning, as 53.6% of AI-driven users bounced immediately.

600% growth, 0.2% share, more than half left directly.

This set of numbers paints a grim picture—AI is siphoning off significant traffic from search engines, but it’s not redirecting that traffic to content creators. It’s keeping the traffic on its own platform.

More intriguing is the behavior of users after clicking from AI to the website.

Multiple analyses indicate that many users arriving from AI chat tools simply want to verify whether the AI’s answers are accurate—they open the webpage, glance at it, confirm the AI isn’t making things up, then close the page and leave. They are not here to “consume content,” but to “verify facts.” This type of traffic has little commercial value for content creators.

This creates a significant "ecological imbalance."

AI companies use content from publishers and content creators to train their models. When users ask AI questions, the AI generates answers using this learned knowledge. Users are satisfied and stay on the AI platform. Advertising revenue (if any) stays with the AI platform. Meanwhile, those who created the original knowledge and content receive almost nothing.

This is not just a matter of traffic allocation. It is undermining the very foundation of the digital content economy.

Over the past 20 years, the fundamental logic of internet digital advertising has been built on the chain of “visit → display → click.” Users visit websites, see ads, and advertisers pay. When the very starting point—“visit”—begins to disappear, the foundation of this entire chain ceases to exist. For media and content platforms that rely on advertising revenue for survival, this is an existential threat.

Gartner predicts that by 2028, organic traffic from traditional search engines will drop by 50%. Not 5%, not 10%—but halved. If this prediction comes true, we will be living in an internet vastly different from today’s.

05 The Deepest Paradox

This may be the deepest paradox of this entire "diversion."

Return to the beginning of Baidu. Baidu is using AI to offset the decline in search advertising—in the first quarter of 2026, AI-driven business revenue reached RMB 13.6 billion, a 49% year-over-year increase, and for the first time accounted for more than half of core business revenue. While Baidu is using AI to dismantle the old Baidu, it is also attempting to rebuild a new Baidu with AI.

But the bigger issue isn’t with Baidu. Baidu at least has the capacity to transform. What about the countless small and medium-sized websites, content platforms, and niche communities that have been sustained by Baidu’s traffic referrals? They lack the ability to develop AI and don’t have the brand recognition to attract users directly. When search engines no longer drive traffic and AI doesn’t either, what will happen to them?

For 20 years, there has been an unwritten "contract" between search engines and websites: you create content, and I give you traffic. This contract has no legal documentation or protection—it’s built on a simple business logic: search engines need high-quality content to attract users, and websites need the traffic from search engines to survive. Both sides benefit.

Now, AI has thrown this equation out of balance. Search engines (and AI platforms) have realized: I don’t need to send users to a website—I can provide the answer myself. Users even prefer this approach—no clicks, no waiting for pages to load, no sifting through ads to find real content.

A product that makes users happier is often one that makes content creators suffer.

When improving user experience and diminishing creators' earnings become two sides of the same coin in an ecosystem, that ecosystem has reached a crossroads requiring fundamental redesign.

AI is killing the ecosystem that feeds it.

No one has yet provided a suitable answer to what the internet will become next.