ChatGPT is still dominant, but dominance is different from exclusivity
Similarweb's July 2026 analysis places ChatGPT at approximately 53% of worldwide generative AI website traffic by May 2026, compared with roughly 76% in June 2025. Over the same period, Gemini rose from below 9% to approximately 27 to 28%, while Claude increased from around 2% to almost 9%. Together, those changes have transformed what was once an overwhelmingly ChatGPT-led category into a much more distributed market.
It is important not to interpret that as evidence that ChatGPT itself is collapsing. Similarweb reports that ChatGPT's absolute website visits remained roughly flat while the overall generative AI category expanded around it. Average monthly visits across generative AI platforms increased by 70% year on year to approximately 9.5 billion between June 2025 and May 2026. ChatGPT owns a smaller percentage of a substantially larger market.
Gemini has become too large for marketers to treat as secondary
The most significant change is arguably Gemini rather than ChatGPT. Moving from less than one tenth of generative AI website traffic to more than one quarter in approximately twelve months means a substantial audience now conducts AI-assisted discovery outside ChatGPT. Gemini's position within Google's wider ecosystem also gives it multiple routes through which users can encounter AI-generated information.
For marketers, this creates a simple problem. A brand can perform extremely well when a potential customer asks ChatGPT for a recommendation and still be missing when the equivalent question is asked elsewhere. If a significant proportion of the addressable audience has moved towards Gemini, knowing only what ChatGPT says about the business provides an increasingly incomplete view of AI visibility.
Claude illustrates how quickly the market can change again
Claude remains much smaller than ChatGPT or Gemini on web traffic, but its trajectory is notable. Similarweb estimates that its worldwide share rose from around 2% to nearly 9% within a year, making it the largest proportional increase among the major platforms in that dataset. The significance is not simply Claude's current market share. It is how quickly an AI platform can move from a relatively small audience to one that marketers can no longer comfortably ignore.
This is why an AI visibility strategy should not be designed around today's market-share table alone. ChatGPT, Gemini, Claude, Perplexity, Copilot, Meta AI and other interfaces do not need equal weighting for every organisation, but businesses should understand which platforms matter to their customers and how that changes over time. Optimising around a single winner assumes the market has already settled. The data suggests the opposite.
There is no universal number-one AI ranking
Google search conditioned marketers to think about rankings as relatively defined positions. A webpage might be first, fifth or twentieth for a particular query. AI recommendations behave differently. ChatGPT can construct one answer, Gemini another and Claude another, even when the user expresses essentially the same need. The platforms use different systems, retrieval processes, sources and response structures.
As a result, being "number one on AI" is not a particularly useful objective. A business may be strongly recommended on one platform, weakly represented on another and completely absent on a third. It may also be recommended for one buyer prompt but not a closely related one. AI visibility is better understood as a pattern across prompts and platforms than as one universal ranking that a company can achieve and defend.
Single-platform AI monitoring is becoming a strategic blind spot
When ChatGPT accounted for around three quarters of generative AI web traffic, concentrating heavily on it was understandable. At approximately 53%, that assumption becomes harder to defend. Similarweb itself concludes that single-platform coverage now misses a meaningful share of the AI-using audience. Gemini alone represents roughly another quarter of worldwide generative AI website traffic in its latest dataset.
This matters most for businesses using manual checks as their AI visibility strategy. Typing three questions into ChatGPT once a month may produce interesting observations, but it does not provide a reliable picture of the market. Commercial prompts need to be tested across the AI platforms relevant to the audience, with competitor recommendations, mentions, citations and changes monitored over time. Without that comparison, a business can appear highly visible simply because it is looking at the platform where it already performs best.
The opportunity is bigger than ChatGPT
The fragmentation of AI traffic should not be viewed negatively. ChatGPT's declining share is occurring because the overall category is expanding. More users are incorporating generative AI into their online behaviour, more platforms are competing for those users and more customer journeys can consequently begin with an AI-generated response. For organisations, the addressable opportunity around AI discovery is becoming larger at the same time as it becomes more complicated.
The businesses that benefit will be those that stop thinking about "ranking in ChatGPT" and start thinking about AI visibility as a channel. That means understanding where customers ask questions, which brands are recommended in response, what sources shape those recommendations and where meaningful gaps exist between platforms. ChatGPT remains essential. It simply no longer represents the entire market, and the gap between those two ideas is becoming more important every month.