a16z Releases Top 100 AI Applications List: ChatGPT Dominance Wanes, Global Markets Split into Three

AI is increasingly deeply embedded in the tools people are already using.

Author: ethn, a16z

Translation: Deep潮 TechFlow

Deep潮 Guide: a16z has released the sixth edition of the generative AI consumer app rankings. ChatGPT has 900 million weekly active users, but Gemini and Claude are growing faster in paid subscriptions. The “default AI assistant” competition has officially begun.

The biggest change in this edition is including long-established products like CapCut, Canva, and Notion—whose “AI features have become core”—in the rankings, along with the first coverage of Agents, AI browsers, and desktop tools.

Author Olivia Moore is a partner in a16z’s consumer team. This report is one of the most systematic publicly available data sources tracking the AI consumer app landscape to date.

Full Text:

Three years ago, we published the first version of this ranking, with a simple goal: to figure out which generative AI products are truly being used by mainstream consumers. At that time, the line between “AI-native” companies and others was clear. ChatGPT, Midjourney, Character.AI were products built from scratch around foundational models, while the rest of the software world was still exploring how to use this technology.

That boundary no longer holds. CapCut is a video editor with 736 million monthly active mobile users, and its most popular features are all AI-driven—background removal, AI effects, auto subtitles, text-to-video. Canva’s growth engine is entirely built on Magic Suite AI tools. Notion’s paid AI adoption rate soared from 20% to over 50% within a year, with AI features now contributing about half of the company’s ARR.

Starting with this edition, we expanded the scope to include all consumer products where generative AI has become a core experience, such as CapCut, Canva, Notion, Picsart, Freepik, and Grammarly. We believe this better reflects how people are actually using AI, though most top-ranked products remain AI-native.

Figure caption: Complete list of the Top 100 Generative AI Consumer Applications as of March 2026

As before, web rankings are based on monthly unique visitors (data from SimilarWeb, as of January 2026), and mobile rankings are based on monthly active users (data from Sensor Tower, as of January 2026). Here are our key findings:

1. ChatGPT leads, but the “Default AI” battle is underway

ChatGPT remains the largest consumer AI product, far ahead of others. Its web traffic is 2.7 times that of the second-place Gemini, and its mobile monthly active users are 2.5 times Gemini’s. ChatGPT’s weekly active user count has grown by 500 million over the past year, now reaching 900 million. Considering that larger scale makes growth harder, this number is remarkable—over 10% of the global population uses ChatGPT weekly.

But we’re seeing the landscape widen, with other general-purpose platforms gaining ground in specific scenarios. Over the past year, paid subscriptions for Gemini and Claude have accelerated in the U.S. (though they still lag far behind ChatGPT—ChatGPT’s paid users are 8 times Claude’s and 4 times Gemini’s). According to Yipit Data, as of January 2026, Claude’s paid user base grew over 200% year-over-year, and Gemini’s by 258%. We also observe increasing multi-platform usage—about 20% of weekly ChatGPT web users also used Gemini in the same week.

What’s happening? Competitors are stepping up. Google has made impressive strides with creative models—Nano Banana generated 200 million images in its first week, bringing 10 million new users to Gemini; Veo 3 is recognized as a breakthrough in AI video. Anthropic is focusing on professional users, launching Cowork, Claude in Chrome, plugins for Excel and PowerPoint, and most notably, Claude Code.

This competition is not just about who is ahead today but about who can build structural barriers. Context accumulates: the more a large language model (LLM) knows about you, the better its results, and the more dependent you become on it. Early data shows Gemini’s web-based per-user monthly conversation count is rising, but ChatGPT still leads by 1.3 times. On mobile, ChatGPT’s advantage is even greater—2.2 times the per-user monthly conversations of Gemini. According to Yipit Data, both companies’ retention rates among U.S. paid consumer users are industry-leading.

The next layer of lock-in comes from app stores. Both ChatGPT and Claude have launched connector ecosystems—ChatGPT with GPTs and Apps, Claude with MCP integrations and Connectors—allowing users to build workflows on top of their assistants. Once users connect AI to their calendars, email, and CRM, switching costs skyrocket. Developers are likely to focus on the platforms with the most users, creating a flywheel similar to early platform wars.

We can already see divergence in platform strategies. Sam Altman has said OpenAI aims “to bring AI to the billions who cannot afford subscriptions,” which is why they are starting to run ads. He also mentioned OpenAI will introduce “Sign in with ChatGPT” identity layer, positioning the AI assistant as the default interface between consumers and the internet. The ambition is for ChatGPT to be the starting point for everything: shopping, hotel bookings, browsing, health management, daily life.

The app directories already reflect this difference. By the end of February, ChatGPT’s app store covered 13 categories with 220 applications. Claude has about 160 curated connectors plus roughly 50 community MCP servers. But only 41 applications overlap—about 11% of the combined catalogs—and these are mostly universal productivity tools everyone needs: Slack, Notion, Figma, Gmail, Google Calendar, HubSpot, Stripe.

Beyond core tools, the two platforms are almost entirely diverging. ChatGPT has over 85 exclusive applications in travel, shopping, food, health, lifestyle, and entertainment—areas where Claude has almost none. These are consumer transactional scenarios: booking flights on Expedia, grocery shopping via Instacart, browsing listings on Zillow, tracking nutrition with MyFitnessPal. This is the most aggressive super-app approach among all AI companies. Claude’s exclusive integrations lean toward professional domains: financial data terminals (PitchBook, FactSet, Moody’s, MSCI), developer infrastructure (Sentry, Supabase, Snowflake, Databricks), scientific and medical tools (PubMed, Clinical Trials, Benchling), and an open-source MCP community without a direct ChatGPT counterpart.

Anthropic seems focused on high-end AI users—developers, knowledge workers—who are more willing and able to pay high subscription fees. While ChatGPT also offers products aimed at similar audiences (like Codex, Frontier), they aim to become truly mainstream platforms—potentially opening more monetization channels as their user base grows. They are already testing ads, and transaction commissions are a natural extension.

If AI assistants evolve from just chat windows to operating system-level environments, this competition might resemble the mobile OS wars—where one platform captures 90% of the market—rather than a search war. Each of these very different platform concepts could build trillion-dollar ecosystems.

2. Global usage is diverging by product

Geographically, the AI market is splitting into three distinct ecosystems, with widening gaps.

Western AI tools share highly similar user bases. The core markets for ChatGPT, Claude, Gemini, and Perplexity are the same: the US, India, Brazil, the UK, and Indonesia—just in different orders. None have significant usage in China or Russia, due to policies: since 2022, Western sanctions restrict Russia’s access to US AI tools; China requires registration, data localization, and censorship compliance.

DeepSeek is the only product crossing multiple camps. Web traffic is distributed across China (33.5%), Russia (7.1%), and the US (6.6%), with similar patterns on mobile. Chinese users also heavily use ByteDance’s Doubao and local products like Kimi.

Russia, which was nearly negligible in our early editions, has now become the third pole, with DeepSeek’s penetration ranking second. Yandex Browser, integrated with Alice AI assistant, has 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 mirrors China’s—just compressed in time: sanctions created a vacuum, and local products filled it within two years.

To measure AI adoption per capita, we built a simple index combining web visits and mobile monthly active users, scaled from 0 to 100. The results redefine the geographic landscape. Singapore ranks first, followed by the UAE, Hong Kong, and South Korea. The US—home to most AI products—ranks 20th.

Figure caption: Generative AI Per Capita Adoption Index (0-100), Singapore first, US 20th

3. Major reshuffle in creative tools

Midjourney, DALL-E, and Stable Diffusion are the products that brought most early users into the generative AI world—three of them launched before ChatGPT. Image generation tools dominated the creative category (video and audio came later), and have consistently ranked high in our top lists. The category has since undergone significant change.

In the first edition in September 2023, 7 of the top 9 creative tools on web were image generators. Three years later, only 3 image generators remain on the list, but there are still 7 creative tools. The difference is what filled the gaps: video, music, and speech products have taken over the space previously occupied by images.

The story of image generation is one of being swallowed by consolidation. As ChatGPT (GPT Image 1.5) and Gemini (Nano Banana) integrated image models with improving quality, the barrier to entry for standalone image products skyrocketed. In our first list, Midjourney ranked in the top 10 but has now fallen to 46th. Remaining products—Leonardo, Ideogram, CivitAI—tend to serve specific creative communities, differentiating with specialized features rather than competing directly with general-purpose generators.

Video generation is the most dynamic area in this edition. Kling AI, Hailuo, and Pixverse have built real user bases, with Chinese-developed models consistently leading in output quality. Applications based on Seedance 2.0 that appear in the next list wouldn’t surprise us. Veo 3 is the first American model to narrow the gap, boosting Google Labs’ traffic (rising from 36th to 25th).

Who’s missing? Sora. OpenAI launched Sora 2.0 as a standalone app in September 2025, allowing users to upload their digital avatars as Cameos to generate real-person videos. Sora topped the US App Store for 20 days straight, reaching 1 million downloads faster than ChatGPT. But downloads later declined because Sora failed to sustain viral growth as a social app (no one has yet cracked the AI + social combo), so it didn’t make the mobile top charts this time. However, SensorTower data shows Sora still has over 3 million daily active users on mobile, and AI video creators continue to use the model, even if they post their work elsewhere.

Music and speech tools are more defensible. Suno maintained its previous ranking (15th). ElevenLabs has been on every edition since September 2023; its core capabilities—voice cloning, dubbing, audio production—are highly professionalized and not yet integrated as a checkbox feature in giant products.

In summary: when giants like Google and OpenAI focus on creative directions (images, increasingly videos), independent products’ traffic gets squeezed—though there’s still room for more niche, attitude-driven, higher-priced products outside the mainstream. Conversely, areas neglected by giants (music, speech) offer larger opportunities for independent innovation.

4. Agents are here

The shift toward agentification of AI didn’t start with this edition—it began earlier, with vibe coding. When Lovable, Cursor, and Bolt appeared on our March 2025 list, they represented a new paradigm: AI products that do more than answer questions or generate media—they build things for users. This is the agent behavior, just limited to specific verticals.

Vibe coding has proven retention among technical and semi-technical users. Both Replit and Lovable are on this list, as is Claude Code (via Claude). There’s more room to grow, as this trend has not yet fully penetrated the mainstream. The top five vibe coding platforms continue to grow, albeit at a slower pace than during their initial explosive phase, and many products are still increasing revenue as developer and team usage deepens.

More recently, general-purpose agents have emerged. In January 2026, an open-source project called OpenClaw went from an independent side project to a GitHub star with 68,000+ stars and media coverage within weeks. Created by Austrian developer Peter Steinberger, OpenClaw is a locally running AI agent that can connect to messaging apps and perform multi-step tasks on your behalf.

If ChatGPT was the moment consumers realized AI could converse, OpenClaw might be the moment they realize AI can act. The product exploded in the developer community—if we push the analysis window to February instead of January, OpenClaw would rank in the top 30 web rankings.

But OpenClaw is not yet a consumer product—it requires terminal skills to install and maintain. It continues to gain momentum among tech-savvy users, becoming GitHub’s most starred project in early March, surpassing React and Linux. However, it has not “graduated” to mainstream users—at least based on new visitor data from its install page, growth remains modest. The project was acquired by OpenAI in February 2026, which may signal a more user-friendly version of OpenClaw is on the way.

OpenClaw isn’t the only general-purpose agent on the list. Manus and Genspark also made the cut—both platforms allow consumers to delegate open-ended tasks (research, spreadsheet analysis, slide creation) to AI, which handles entire workflows end-to-end. Manus is making its second appearance; after debuting on the list, it was acquired by Meta in December 2025 for about $2 billion. Genspark is a new entrant—earlier this year, it completed a $300 million Series B and announced $100 million annualized revenue.

On mobile, consumers typically interact with agents via text rather than dedicated apps. During installation, users connect OpenClaw to platforms like WhatsApp, Telegram, and Signal, giving commands as if messaging a friend, with tasks executed in the background. Other products like Poke also offer similar agent experiences via SMS.

These products will directly compete with the agent capabilities of the major consumer LLMs—ChatGPT, Claude, Gemini. As these giants build their own connector and app ecosystems, will consumers choose one as their primary agent? The next six months will tell.

5. AI moves beyond browsers and apps

Previous editions ranked AI products by web traffic and mobile monthly active users. But a new class of AI products is emerging that these metrics can’t capture. Over the past year, some of the most significant consumer AI growth has occurred in products invisible to these two indicators.

The most obvious change is that browsers themselves are becoming AI products. Over the past nine months, OpenAI launched Atlas—a browser with ChatGPT embedded in every page—Perplexity introduced Comet, and Browser Company (later acquired by Atlassian) launched Dia. Yipit data shows that Perplexity’s Comet has made the biggest impact (measured by downloads), but no AI browser has yet achieved accelerated growth.

Other AI giants are integrating AI into existing browsers rather than launching standalone AI browsers. Google added Gemini to Chrome and released a beta called Disco, which can dynamically generate web apps based on your browser tabs. Anthropic released Claude in Chrome, allowing connection to Claude or Claude Code sessions to drive web-based actions.

Growth of native desktop AI tools is even more rapid, especially for developer tools. Claude Code—a command-line developer agent—reached $1 billion annualized revenue in just six months. OpenAI launched a Mac-native Codex app, which the company says had 2 million weekly active users in early March, up 25% week-over-week. Cursor remains in the top 50 web rankings.

For pure consumers, the most common standalone desktop AI applications are speech-related. Tools like Fireflies, Fathom, Otter, TL;DV, and Granola, which focus on note-taking, reach users via PLG models and are gradually penetrating enterprises—together, the top five have over 20 million visitors. Workspace apps like Notion (first on this list) increasingly integrate AI through note-taking, research agents, and task automation.

Finally, AI is embedding more deeply into existing tools people already use. Anthropic launched Claude in Excel and PowerPoint. OpenAI introduced ChatGPT for Excel. Google deepened Gemini’s integration across Workspace—Docs, Sheets, Gmail, Meet—all now have native AI features. In January 2026, Google launched Personal Intelligence, connecting Gemini to Gmail, Google Photos, YouTube, and Search, enabling the assistant to reference your hotel bookings, purchase history, albums, and viewing history without you explicitly telling it.

The takeaway from this list: we are underestimating the most-used AI products. A developer spending eight hours daily in Claude Code, or a knowledge worker dictating every email via Wispr, are heavy AI users but almost invisible in web traffic data. When AI shifts from being a destination to a core function, our methodology must evolve accordingly.