The Big Winners and Losers in U.S. Website Traffic Over the Past Year

We often talk about the "AI revolution" in search as a theoretical shift or a series of scary headlines. But when you stop looking at the hype and start looking at the actual movement of billions of users, a different story emerges. The way people navigate the internet is changing, and the data shows that the gap between the winners and everyone else is widening.

The problem is that we are trying to measure this shift while our primary tools are in flux. Between the forced migration to GA4 and the rollout of AI Overviews, we lost our historical baselines exactly when we needed them most. To get a real sense of where the U.S. audience is moving, we have to look at third party data that doesn't rely on the "vibes" of a dashboard.

The giants are growing, not just staying put

If you look at a list of the most visited sites in the U.S., the names at the top look familiar. Google is still number one with 25.3 billion monthly visits, followed by YouTube at 10.3 billion and Reddit at 2.61 billion. At a glance, it looks like the status quo has held. But the actual volume tells a different story.

Google's traffic has climbed by 10.72% year over year, and YouTube has seen a massive jump of 36.6%. These aren't just incremental gains; they are a consolidation of power. The top three sites now command 54.3% of all traffic within the top 10. While the total visits across the top 150 U.S. sites grew by a modest 6.1%, the top 10 still hold 68.6% of that total share. The same pattern also shows up in AI Search Traffic Differs from Organic Traffic, where the practical question is how the signal becomes visible.

The takeaway here is that the web is not becoming more decentralized. If anything, the concentration of attention is tightening. The "big players" aren't just surviving the AI transition; they are capturing more of the total available attention.

Expert Interpretation: For a business owner, this means the "long tail" of the web is shrinking. When traffic concentrates at the top, the cost of acquiring a user from those platforms increases. The tradeoff is simplicity versus reach. You can still reach people, but you are increasingly dependent on a handful of gatekeepers who control the majority of the U.S. web's attention.

ChatGPT moves from novelty to habit

The most significant shift in the landscape is the rise of ChatGPT. It has climbed to the ninth most visited website in the U.S., pulling in 1.09 billion monthly visits. That is a year over year increase of 48.38%.

What is striking is that an AI chatbot now outranks established search engines like Bing and DuckDuckGo. This suggests that generative AI has moved past the "experimentation" phase and has become a daily habit for a billion people. Interestingly, this isn't a general win for all new platforms. TikTok, for example, ranks 14th with 573.6 million visits, but its growth was a modest 1.63%.

There are other non AI winners as well. LinkedIn moved up to 19th place with a 15.48% increase in visits. USPS.com also saw a 12.04% bump, proving that basic utility and habit still drive massive traffic. Twitch also saw a jump, climbing to 25th place after a 25.6% increase in visits, and Comix.to saw the largest percentage gain of all, jumping 14 spots to 56th with a 37.3% increase.

Expert Interpretation: The surge in ChatGPT traffic indicates a fundamental change in user intent. People are no longer just searching for a list of links; they are seeking direct answers. The decision you need to inspect here is your content's "answerability." If your value proposition is simply providing a list of facts that an AI can summarize, you are competing with a site that is growing by nearly 50% a year. This connects with Not the Outputs) when the same signal needs a clearer operating decision.

The surprising decline of the "AI powered" search

The losers in this data set are perhaps more telling than the winners. The steepest decline in the top 20 belongs to Bing, which plummeted 50.43% year over year. This is a surprising result considering the massive investment Microsoft made to integrate Copilot into Bing to steal search share from Google.

The data suggests that bolting AI onto a search engine didn't necessarily create a new habit for users. Bing's standalone search share has remained largely flat against Google for years, and this traffic drop suggests the AI hype didn't translate into long term user retention. A useful companion note is 000x Human Traffic in 5 Years, because it looks at a nearby part of the same system.

Temu also saw a significant drop of 28.55%. However, this wasn't an AI or SEO failure. It was a result of trade policy. The closure of the "de minimis" tariff exemption for low value Chinese imports in mid 2025 removed the pricing advantage that fueled Temu's growth. This serves as a reminder that external regulatory shifts can impact traffic just as heavily as algorithm updates.

Other notable losers include Amazon, down 15.31%, and DuckDuckGo, which fell 15.83%. For those of us who view Amazon as the bellwether for ecommerce health, a 15% drop in traffic is a signal that warrants a closer look at consumer behavior.

Expert Interpretation: Bing's failure to capture the AI wave despite having the technology first proves that "features" do not equal "distribution." The tradeoff here is between technical capability and user habit. You can have the best AI tool in the world, but if the user's habit is rooted elsewhere, the tool alone won't move the needle. Inspect whether your growth strategy relies on a "better feature" or a "better habit."

The measurement gap in GA4

This is where the situation becomes problematic for SEOs. There is a growing gap between what is actually happening on the web and what our analytics tools are telling us. Many of us are relying on GA4, but GA4 is increasingly blind to the sources of our traffic.

Consider a niche publishing site where the GA4 reports look fantastic: sessions are up 105%, engaged sessions are up 111%, and key events are up 80.5%. On the surface, it looks like a massive win. But when you dig deeper, you see that "direct" traffic has surged by 157% and now accounts for 78% of all traffic.

The reality is that "direct" is not a source. It is a bucket for everything GA4 cannot identify. A significant portion of this "direct" traffic is likely coming from AI chat interfaces. When a user clicks a link inside an AI response, the referrer data is often stripped away, leaving GA4 to label the visit as "direct."

Expert Interpretation: We are currently in a "Bermuda Triangle" of analytics. If you see a spike in direct traffic, you cannot assume it is brand loyalty or bookmarks. It could be an AI recommendation. The danger is making strategic decisions based on a "win" that you cannot actually attribute or replicate because you don't know where it's coming from.

Ignoring the noise of default channels

In the movie *The Gumball Rally*, a racer rips the rearview mirror off his car, claiming that what is behind him is not important. In the current state of web analytics, we need to apply a similar logic to GA4's default channels.

GA4 provides 19 default channels, including a specific "AI Assistant" channel designed to track visits from Gemini, ChatGPT, Copilot, and others. However, this channel is largely useless. It excludes Google's own AI Overviews and AI Mode, and in the case of the niche publisher mentioned earlier, it only accounted for 0.47% of total traffic.

The traffic is arriving, but it isn't showing up in the "AI Assistant" bucket. It is hiding in the "direct" bucket. Relying on the default labels provided by the tool is giving us a false sense of security or a false sense of failure.

Expert Interpretation: The tradeoff here is between convenience and accuracy. It is convenient to trust the default dashboard, but it is inaccurate. The decision you must make is to stop treating GA4 as a source of truth for *where* traffic comes from and start treating it as a tool for *what* users do once they arrive.

Expanding your intelligence toolkit

If the primary tool for SEOs is becoming less capable of attributing traffic, the only way to survive is to diversify how you measure success. We cannot rely on a single dashboard to tell us if we are winning or losing in the age of AI.

The data shows that while Google and YouTube continue to dominate, and ChatGPT is rapidly becoming a primary destination, the "middle class" of the web is struggling. To avoid becoming a cautionary tale, you have to look beyond the default channels of your analytics software.

This means investing in audience intelligence and market research tools that can see the patterns GA4 misses. You need to understand the ecosystem from the outside in, rather than trying to deduce the world from a stripped down referrer report. The winners of the next year will be those who stop looking in the rearview mirror of traditional analytics and start using tools that provide a wider field of vision.

Expert Interpretation: The ultimate risk is "invisible decline." You might see your traffic holding steady in GA4, but you could be losing your presence in the AI driven discovery layer. The only way to mitigate this is to use third party traffic analytics and survey tools to validate your internal data. If your internal data says you are winning, but market research shows your competitors are the ones being recommended by AI, your GA4 report is lying to you.