New data from Similarweb highlights how users are accessing the leading generative AI tools, revealing a clear and consistent preference for desktop use. Despite the widespread adoption of smartphones, the majority of interactions with large AI models still occur on traditional computers, suggesting that generative AI remains a predominantly desktop-driven activity.
Across the eight major GenAI platforms analyzed, desktop traffic significantly outweighs mobile traffic. Claude stands out with the highest proportion of desktop usage—88.9 percent of its traffic originates from desktop devices, leaving just 11.1 percent from mobile. Grok, DeepSeek, and Perplexity all show a similar pattern, with around 83 percent of their users accessing the tools from desktop and roughly 17 percent from mobile. Even more consumer-oriented platforms such as ChatGPT and Gemini show a clear desktop bias, with 70.0 percent and 69.4 percent of their traffic coming from desktop respectively, though their mobile usage is relatively higher than the others. Microsoft Copilot sits close behind with 68.9 percent of its traffic from desktop, while Meta AI demonstrates the most balanced distribution—74.3 percent desktop and 25.7 percent mobile.
The dominance of desktop traffic can be attributed to several overlapping factors. A large portion of these tools are used within professional and academic environments, where desktops and laptops remain the primary devices. AI models like Claude, Perplexity, and Microsoft Copilot are often integrated into workflows centered around productivity, research, and writing, making them more naturally suited to desktop use. In particular, Microsoft Copilot’s deep integration with Microsoft 365 applications reinforces this tendency, as these tools are primarily desktop-first in nature.
Another explanation lies in the complexity of the tasks these models are used for. Many generative AI interactions involve coding, data analysis, or the generation of long-form content—activities that are difficult to perform efficiently on mobile devices. Desktop interfaces offer larger screens, easier text manipulation, and the ability to multitask across multiple windows and tools. This technical and ergonomic advantage helps explain why even consumer-accessible models like ChatGPT still see a strong majority of their traffic from desktops.
Interface design also plays a role. While tools such as ChatGPT and Gemini have invested heavily in developing mobile applications, many AI platforms still rely on web interfaces optimized for desktop use. Even where mobile apps exist, users may prefer the richer, more stable desktop experience for extended sessions, file uploads, or professional writing and analysis.
Enterprise adoption further reinforces this pattern. Corporate environments tend to rely heavily on desktop systems for security, productivity, and compliance reasons, meaning that enterprise-integrated tools like Claude, Copilot, and DeepSeek naturally see most of their traffic from desktop users. Generative AI is, in many cases, becoming embedded within professional software suites rather than being accessed independently through mobile devices.
Still, mobile traffic is gradually gaining importance. Around one in four ChatGPT or Meta AI sessions now occurs on mobile, reflecting a shift toward more casual and conversational use. Meta’s relatively high share of mobile users likely stems from its integration within popular social platforms such as Facebook, Instagram, and WhatsApp, where AI interactions happen more spontaneously. OpenAI’s official ChatGPT mobile app has also improved the mobile user experience, encouraging people to use AI on the go for quick questions, creative brainstorming, or conversational assistance. ChatGPT and Gemini also have popular image generation features, which are likely used by casual users on mobile phones.
Overall, the data reinforces that generative AI remains primarily a desktop-first phenomenon, driven by professional, technical, and productivity-oriented use cases. However, as mobile interfaces become more refined and AI becomes increasingly embedded in everyday applications, the balance may begin to shift. The next phase of growth for generative AI could very well be defined by the move from the desktop workspace to the mobile pocket, as companies like OpenAI, Google, and Meta push toward making AI accessible wherever users are.