a16z: Top 100 Global Generative AI Consumer Applications in March 2026

Author: Olivia Moore, Partner at a16z; Translation: Shaw, Jinse FinanceThree years ago, we released the first version of this list with a simple goal: to find out which generative artificial intelligence (AI) products are truly being used by mainstream consumers. At the time, the lines between "AI-first" companies and other businesses were very clear. ChatGPT, Midjourney, and Character.AI were all products built around a foundational model, while other software companies were still figuring out how to apply the technology. This boundary no longer exists.. CapCut, a video editing app with 736 million monthly active mobile users, relies on AI technology for its most popular features—background removal, AI effects, automatic captions, and text-to-video conversion. Canva, a design platform, builds its growth on its AI toolkit, "Magic Suite." Notion's AI-driven paid subscription rate surged from 20% to over 50% in a year, and AI features now account for about half of the company's annual recurring revenue (ARR). Starting with this edition of the list, we will expand our scope to include all consumer products where generative AI has become a core experience, including CapCut, Canva, Notion, Picsart, Freepik, and Grammarly. We believe this will more accurately reflect how people actually use AI, although the top products on the list are still predominantly native AI products. As always, our website rankings are based on monthly unique visitors from SimilarWeb as of January 2026. Mobile apps are ranked based on monthly active users from Sensor Tower as of January 2026. Here are our core findings:1. ChatGPT remains in the lead, but the battle for "default AI" has begunChatGPT remains the largest consumer AI product by a wide margin. On the web, its monthly visits are 2.7 times that of Gemini, which ranks second; on mobile, its monthly active users are 2.5 times that of Gemini. Over the past year, ChatGPT's weekly active users have increased by 500 million, reaching 900 million today. This achievement is particularly impressive considering the difficulty of maintaining growth at such a massive scale—currently, over 10% of the global population uses ChatGPT weekly. But we are beginning to see that this area is expanding, and other general-purpose platforms are gradually emerging in specific application scenarios. Over the past year, both Gemini and Claude have seen accelerated growth in paid subscribers in the US (though still far behind ChatGPT – by this metric, ChatGPT is 8 times the size of Claude and 4 times the size of Gemini). According to Yipit Data data as of January 2026, Claude's paid subscriber growth exceeded 200% year-over-year, while Gemini grew by 258%. Furthermore, the phenomenon of using multiple platforms simultaneously is increasing – approximately 20% of weekly ChatGPT web users also use Gemini within the same week. What exactly happened? Competitors have stepped up their efforts. Google has performed strongly in the creative model field—Nano Banana generated 200 million images in its first week, bringing 10 million new users to Gemini; while Veo 3 is widely regarded as a breakthrough moment in the field of AI video. Anthropic continues to increase its efforts for high-end individual and light business users, launching Cowork, Chrome version of Claude, and Excel and PowerPoint plugins, among which Claude Code is the most noteworthy. The significance of this competition lies not only in who can currently lead, but also in who can establish an unshakeable structural advantage. Contextual data has a cumulative effect: the more a Large Language Model (LLM) knows about you, the better the results it provides, and the more frequently you will use it. Early data shows that Gemini's monthly usage on the web is climbing, but it is still 30% lower than ChatGPT (ChatGPT is 1.3 times higher); while on mobile, ChatGPT is absolutely dominant, with monthly usage 2.2 times that of Gemini. According to Yipit Data, both companies have top-tier consumer-paying user retention rates in the US market. Next layer user lock from app store. Both ChatGPT and Claude have launched connected ecosystems—ChatGPT's custom GPTs and applications, and Claude's MCP integrations and connectors—allowing users to build workflows on top of their AI assistants. Once users configure their AI to interface with calendars, email, and customer relationship management (CRM) systems, the switching costs rise sharply. Developers may focus their efforts on the platforms with the largest user base, creating a flywheel effect similar to the platform wars of the past. We can already see the development direction of each platform. Sam Altman previously stated that OpenAI hopes to "bring AI to billions of users who cannot afford subscriptions," which is why they launched advertising; he also revealed that OpenAI will launch an identity layer of "login with ChatGPT," positioning the assistant as the default interaction point between users and the internet. The goal is to make ChatGPT the starting point for everything: shopping, booking, browsing, health management, and daily life. The application catalog already reflects this. As of late February, ChatGPT's app store covered 13 major categories and 220 apps. Claude had approximately 160 curated connectors, plus about 50 community-hosted MCP servers. However, there was only 41 overlap between the two apps, accounting for about 11% of the total number of apps, and these 41 were almost all general productivity tools: Slack, Notion, Figma, Gmail, Google Calendar, HubSpot, and Stripe. Aside from core tools, the two platforms' ecosystems are almost completely divergent. ChatGPT has over 85 apps covering travel, shopping, food, health and wellness, lifestyle, and entertainment, while Claude has almost none in these categories. These are all consumer transaction scenarios: booking flights on Expedia, ordering groceries through Instacart, browsing accommodations on Zillow, and tracking nutrition intake on MyFitnessPal. This is the most aggressive consumer-grade super-application strategy among all AI companies. Claude's exclusive integrations, on the other hand, lean towards professional fields: financial data terminals (PitchBook, FactSet, Moody's, MSCI), developer infrastructure (Sentry, Supabase, Snowflake, Databricks), scientific and medical tools (PubMed, clinical trial databases, Benchling), and the ever-growing open-source MCP community—something ChatGPT lacks. Anthropic seems to focus on heavy AI users (developers, knowledge workers, etc.). These users are currently more willing and able to pay for expensive direct subscription services. While ChatGPT also has products targeting the same group (such as Codex and Frontier), they have also expressed a vision to become a truly mainstream user platform – which could open up more monetization avenues as their user base continues to expand. They are already testing advertising, and charging commissions on transactions would be a natural expansion direction. If AI assistants are no longer just chat windows but become an operating environment, this competition may not ultimately resemble the search engine wars, where one company holds 90% of the market share, but rather a mobile operating system war: two platforms with vastly different philosophies, each building a trillion-dollar ecosystem. 2. Global Usage is Differentiating by Product Geographically, the AI market is splitting into three distinct ecosystems, and the gap between them is widening. The user groups of Western AI tools still highly overlap. ChatGPT, Claude, Gemini, and Perplexity all rank highly in the same group of countries: the United States, India, Brazil, the United Kingdom, and Indonesia, only in different orders. These products have not achieved significant usage in China and Russia. This is a result of policy: since 2022, Western technology sanctions have restricted Russia's use of US AI tools; while China requires AI service providers to register, store data domestically, and comply with content review regulations. DeepSeek is the only product to cross this barrier. Its web traffic distribution is: China (33.5%), Russia (7.1%), and the United States (6.6%), with a similar pattern on mobile. Chinese users also heavily utilize ByteDance's Doubao and the local product Kimi. Russia, which was almost invisible in our earlier rankings, has now become our third largest core region, with DeepSeek ranking second globally in penetration there. Yandex Browser, integrated with the Alice AI assistant, boasts 71 million monthly active users, ranking among the top ten global mobile AI products. Sber's GigaChat also appears on our web rankings for the first time. This pattern is similar to that in China, but the development has been much faster: sanctions created a market gap, which was filled by local products within two years. To observe the adoption of artificial intelligence from a per capita perspective, we constructed a simplified index that combines the average number of webpage views per user and the average number of monthly active mobile users per user across all products on the list, with a score range of 0-100. The results reshaped the regional landscape: Singapore ranked first, followed by the UAE, Hong Kong, and South Korea. The United States, where the vast majority of related products originated, ranked only 20th. 3. Creative Tools Play a RoleMidjourney, DALLE, and Stable Diffusion were the products that introduced most early users to generative AI—all three were launched before ChatGPT was released. Image generation tools not only dominated the creative tools space (video and audio generation came later), but also held absolute dominance in our first three lists. Since then, this category has changed dramatically. In the first ranking in September 2023, 7 out of 9 creative tools listed on the web were image generators. Three years later, only 3 image generators remained on the list, but the number of creative tools still reached 7. The difference lies in the fact that video, music, and voice products filled the gaps, taking the places vacated by image generators. The story of image-based products is essentially about ecosystem integration. With the continuous upgrades to ChatGPT's built-in native image model (GPT Image 1.5) and Gemini (nano banana), the barrier to entry for independent image generation products has been significantly raised. In the first edition of this list, Midjourney was in the top ten, but has now fallen to 46th. The remaining products—Leonardo, Ideogram, and CivitAI—are mostly geared towards specific creative communities, offering distinctive features, rather than directly competing in the general generation field. The most significant changes in this edition's list are in the video generation field. Kling AI, Hailuo, and Pixverse have all achieved real user growth, and models developed in China continue to lead in generation quality—we have every reason to expect applications based on Seedance 2.0 to appear in the next edition's list. Veo 3 is the first US model to close this gap and has brought traffic to Google Labs (its ranking rose from 36th to 25th). What's missing? Sora. In September 2025, OpenAI launched its flagship video model version 2.0 as a standalone app, allowing users to upload their digital avatars as "digital substitutes" to generate videos featuring real people. Sora stayed at the top of the US app store charts for 20 days, surpassing one million downloads faster than ChatGPT. Afterward, downloads declined as Sora no longer experienced the viral explosion of social products (currently, no one has truly integrated "AI + social"), thus failing to enter this period's mobile charts. However, according to SensorTower, Sora still has over 3 million daily active users on mobile—even as AI video creators distribute their work across platforms, they continue to use this model. The competitive advantage in the music and voice fields is even more solid. Suno (ranked 15th) maintained its position from the previous ranking. ElevenLabs has been on the list every month since September 2023, demonstrating its professional capabilities in voice cloning, dubbing, and audio production, which have not yet been simply copied as a "standard feature" by giants. The pattern is this: in creative fields (images, and increasingly, videos) where model giants like Google and OpenAI are focusing their efforts, the traffic for independent products is constantly being squeezed. However, products with clearer positioning, potentially higher commercialization, and serving users outside the mainstream still have room to survive. In areas where they haven't yet heavily invested (music, voice), independent products have even more opportunities. 4. The Arrival of the AI Agnet Era The transformation of artificial intelligence into intelligent agents (AI Agnets) didn't begin with this list, but rather started in the previous issue's vibe coding section. When Lovable, Cursor, and Bolt appeared on our March 2025 list, they represented a completely new form: AI products no longer just answer questions or generate media content, but directly create and build for users. This is the behavior of intelligent agents in vertical scenarios. Vibe coding has proven to have a high retention rate among tech and semi-technical users. This week's list includes Replit, Lovable, and Claude Code (based on Claude). This trend has not yet truly penetrated the mass market, and there is still greater room for growth in the future. Traffic to the top five ambient coding platforms is still growing, although the initial explosive growth has slowed; however, as developers and teams increase their usage, the revenue of many of these products continues to rise. Recently, **general-purpose agents** have begun to emerge. In January 2026, an open-source project called **OpenClaw** garnered 68,000 GitHub stars and mainstream media coverage in just a few weeks, transforming from a developer's side project. Developed by Austrian developer Peter Steinberger, OpenClaw is a locally running AI agent that can integrate with your instant messaging software and perform multi-step tasks for you. If ChatGPT introduced consumers to the idea that AI could chat, then OpenClaw may be introducing them to the idea that AI can act. The product quickly gained popularity in the developer community—if our statistics were updated to February (instead of January), OpenClaw would be among the top 30 on our webpage. However, OpenClaw is (currently) not a consumer product; it requires terminal knowledge for installation and maintenance. OpenClaw continues to gain popularity among tech users, becoming the most-starred project on GitHub in early March, surpassing React and Linux. However, the product has not yet truly reached the general public—at least judging from the relatively stable number of new visits to the OpenClaw installation page. The product was acquired by OpenAI in February 2026, which may mean that a more user-friendly version of OpenClaw is coming soon. OpenClaw is not the only general-purpose agent on the list. Manus and Genspark also made the list. Both platforms allow users to delegate open-ended tasks (research, spreadsheet analysis, PPT generation) to AI, which then completes the entire workflow end-to-end. This is Manus's second time on the list; after its initial appearance, it was acquired by Meta in December 2025 for a valuation of approximately $2 billion. Genspark, on the other hand, is a first-time entrant: the company completed a $300 million Series B funding round earlier this year and announced annualized revenue of $100 million. On mobile devices, users typically interact with the agent via text rather than a standalone app. During setup, users can connect OpenClaw to platforms like WhatsApp, Telegram, and Signal; sending it a message, like chatting with a friend, will allow it to perform tasks in the background. Other products like Poke offer similar agent experiences directly via SMS. These products will compete with the intelligent agent functions of general-purpose large-model assistants used by users daily, such as ChatGPT, Claude, and Gemini. As they build their own ecosystems through connectors and applications, will users use one of these products as their primary intelligent agent? The next six months will provide a clear answer. 5. AI Products Beyond Browsers and Apps This list has previously ranked AI products using two metrics: **web page views and mobile monthly active users**. However, a new category of AI products is emerging that cannot be covered by existing ranking metrics. The most significant growth in the consumer AI field over the past year has occurred in products for which neither of these two metrics can be statistically analyzed. The most obvious change is that **browsers themselves are becoming AI products**. Over the past nine months, OpenAI launched Atlas (a browser with ChatGPT built into every page), Perplexity released Comet, and The Browser Company, acquired by Atlassian, launched Dia. Yipit data shows that Perplexity's Comet has had the biggest impact on the market (in terms of download page views), but so far no AI browser has achieved explosive growth. Other AI giants have chosen to add AI functionality to existing browsers rather than launching standalone AI browsers. Google integrated Gemini into Chrome and released Disco, currently in beta—which dynamically generates web applications based on the user's browser tabs. Anthropic launched the Claude plugin for Chrome, which can connect to the user's Claude or Claude Code session and perform actions on web pages. Native desktop AI tools have seen even more rapid growth, with developer tools being particularly prominent. Claude Code, a command-line AI agent for developers, achieved $1 billion in annualized revenue in just six months. OpenAI launched a standalone Mac application, Codex, which, as of early March, had 2 million weekly active users, with a weekly growth rate of 25%. Furthermore, Cursor maintained its position in the global top 50 webpage rankings. For the average consumer, the most common standalone AI desktop applications are mostly voice-related. Note-taking tools such as Fireflies, Fathom, Otter, TL;DV, and Granola reach users through a product-driven growth (PLG) model and are gradually penetrating the enterprise market; the top five products have a combined visitor count of 20 million. Office collaboration applications such as Notion (making its debut on the list this period) are also continuously deepening AI integration through features such as note-taking assistants, research agents, and even task automation. Finally, AI is increasingly being embedded in the tools people are already using. Anthropic launched Claude for Excel and Claude for PowerPoint; OpenAI released ChatGPT for Excel; and Google has deeply integrated Gemini into its entire Workspace suite—Docs, Sheets, Gmail, and Meet now all have native AI capabilities. Google also launched Personal Intelligence in January 2026, integrating Gemini with Gmail, Google Photos, YouTube, and Search, allowing the assistant to access your hotel bookings, spending records, photo albums, and viewing habits without additional commands. The significance of this list is that our rankings are increasingly underestimating the AI products people actually use most frequently. A developer who spends 8 hours a day on Claude Code, a knowledge worker who dictates all their emails via Wispr, are both heavy AI users, but their behavior is rarely reflected in web traffic data. As AI moves from standalone products to built-in features, our statistical methods need to be adjusted accordingly.