ChatGPT's Monopoly Is Melting: 22 Points of Market Share Lost in 14 Months

When ChatGPT launched in November 2022, it had no meaningful competition. When it crossed 100 million users in early 2023, it still had no meaningful competition. Even through most of 2024, as rivals multiplied, ChatGPT was the AI chatbot market in a way that Google once was the search market.

That era is ending.

A new report from Momentic tracking web traffic across the seven largest generative AI chatbots shows ChatGPT’s worldwide share of visits has dropped from approximately 76.5% in February 2025 to 54.7% in April 2026 — a decline of nearly 22 percentage points in fourteen months. The company still leads by a wide margin. But the shape of the market has changed fundamentally, and the trajectory favors challengers.

The Numbers

ChatGPT drew 5.51 billion web visits in April 2026, a number that would have seemed science fiction three years ago. But its share of measured AI chatbot traffic — across platforms including Gemini, Claude, Copilot, Grok, Perplexity, and others — has contracted significantly.

Google Gemini has become the most consequential challenger. Now sitting at 27.4% of measured traffic with 2.76 billion visits, Gemini has grown approximately 104% over the past six months, fueled by Google’s ability to embed the assistant across Search, Chrome, Android, Workspace, and the broader ecosystem that touches billions of users daily. Gemini’s trajectory is not primarily about winning users away from ChatGPT — it is about activating Google’s existing user base at scale.

Claude presents the most intriguing statistical story. Anthropic’s assistant holds just 8.2% of worldwide web visits at 824 million monthly, third place overall. But the growth rate is remarkable: Claude surged 306% in a single quarter, from 203 million web visits in January 2026 to 824 million in April 2026. Over six months, the growth rate is 357%. Claude’s domestic performance is stronger, with a 12.5% share in the U.S. market, where it is increasingly used by developers, knowledge workers, and enterprise teams rather than casual users.

What the Numbers Exclude — and Why It Matters

Before drawing sweeping conclusions, it is worth understanding what Momentic’s methodology measures and what it does not.

The report tracks consumer web traffic via Similarweb’s panel methodology — essentially, the desktop browser visits that flow to chatbot websites. It explicitly excludes native mobile apps, which is where a substantial fraction of ChatGPT’s usage now occurs. It excludes embedded surfaces: Windows Copilot, Android Gemini, the Gemini integrations baked into Google Workspace, and Apple Intelligence’s Siri. It excludes API usage, where enterprise developers are paying to run models at scale for their own products.

ChatGPT’s web visit decline of roughly 6% over six months does not mean fewer people are using it — the company’s own announcement that it crossed one billion weekly active users likely reflects growth across all surfaces, including apps and API. What the web traffic data captures is a specific kind of casual, direct-browser-visit usage, and within that category, the competition is intensifying.

Claude’s 306% quarterly growth, measured on the same methodology, is also worth contextualizing. Growing from 203 million to 824 million web visits is extraordinary, but the absolute numbers still trail ChatGPT and Gemini by a significant margin. Claude is growing faster but from a much smaller base.

Why ChatGPT Is Losing Web Share

The more interesting question is not whether these numbers are real but why they are moving in this direction.

Google Gemini’s rise is almost entirely explained by distribution. Google has quietly inserted Gemini into nearly every consumer touchpoint it controls, and as users encounter it as the default AI layer in Search’s AI Mode, they begin to treat it as a primary assistant rather than a novelty. The 27.4% share reflects that integration more than any particular model advantage.

Claude’s growth tells a different story about engagement quality. Momentic’s data shows Claude with a 26% bounce rate — the lowest among the top AI chatbots measured — and 4.34 pages per visit. These metrics suggest Claude users are not accidentally landing on the site and leaving; they are coming with specific intent and staying to complete multi-step tasks. The pattern aligns with Anthropic’s positioning: a less casual, higher-fidelity assistant that performs better on complex professional tasks, attracting users who want deeper work rather than quick answers.

ChatGPT’s web traffic profile shows a platform facing the classic challenge of the incumbent: the user base is massive but increasingly diffuse. Reaching one billion weekly active users is an extraordinary achievement, but serving that audience means serving everyone from students asking homework questions to executives doing strategic analysis. No single product excels at the entire range, and the specific user segments — developers, enterprise knowledge workers, and sophisticated AI-native users — are precisely those who are beginning to find specialized alternatives more compelling.

The Grok Anomaly

Grok, xAI’s assistant embedded in X (formerly Twitter), shows engagement metrics that Momentic’s researchers flagged as potentially anomalous: 11-minute average session length and 13 pages per visit, far exceeding any other platform. The researchers note these figures may be inflated by Grok’s deep integration into X’s interface, where users encounter it during normal browsing activity that the methodology records as chatbot usage. Grok’s market share remains small in absolute terms, but the engagement pattern — whether authentic or artifact — is worth watching.

The Competitive Landscape at Midyear

The structure of the AI chatbot market at mid-2026 looks less like a monopoly and more like an early-stage multi-player market settling into tiers. ChatGPT is the clear volume leader, with consumer awareness and app-store dominance no competitor comes close to matching. Gemini is the distribution play, winning by embedding into existing Google surfaces rather than persuading users to adopt a new product. Claude is the quality leader among knowledge workers, with the deepest engagement metrics despite the smallest traffic base. Grok is the X-native play, with real uncertainty about its broader relevance.

What the market does not yet have is a single point where the competitive dynamics resolve. ChatGPT’s lead is large enough that OpenAI has enormous room to respond, and the company has been releasing new capabilities at an aggressive pace — from GPT-5.5 to the Codex enterprise platform to the ongoing expansion of ChatGPT’s memory and personalization systems. Google has nearly unlimited distribution leverage to push Gemini deeper. Anthropic’s Claude is gaining ground in enterprise accounts and developer ecosystems where it increasingly competes as a premium alternative.

The more relevant question is not whether ChatGPT maintains 54.7% share or falls to 45% over the next twelve months. It is whether the market evolves toward a model where different AI systems serve different use cases — just as the internet eventually produced specialized tools for search, social, commerce, and communication, each with its own dominant player. That stratification is already visible in the engagement data, and it is where the most interesting competitive battles of the next two years will be fought.

For now, the report’s central finding stands: the AI market that one company once owned is now genuinely contested, and the speed of that change — most of it occurring in a fourteen-month window — is itself the most important data point for anyone building, investing, or competing in the space.