The Top 100 GenAI Consumer Apps

Generative AI 100 List Methodology & Cadence

• The Consumer Team plans to publish the "Generative AI 100" list every six months, with a potential acceleration to every three months due to the sector's velocity.
• Data sourcing involves screening tens of thousands of global websites via SimilarWeb and app store data via Sensor Tower to identify top monthly visitors/users.
• The selection process filters tens of thousands of data points to isolate the top 50 AI companies for web and the top 50 for mobile.

Divergence Between Web and Mobile Ecosystems

• There is minimal overlap between lists, with only five to six companies appearing on both the top web and mobile rankings.
• Mobile-Specific Trends:
  • EdTech: Apps like ELSA leverage mobile audio for real-time pronunciation feedback and photo-based homework helpers utilize the camera.
  • Avatar & Selfie Tools: Products like Lenza and Epic capitalize on existing camera rolls to generate avatars, with users willing to pay approximately $20 for these services.
  • Niche Utility: Keyboard integration tools analyze screenshots (e.g., dating app chats) to suggest responses to increase user charisma.
• App Store Arbitrage:
  • Many mobile apps are "copycat" products using slightly different names to arbitrage App Store Optimization (ASO) keywords after OpenAI was slow to launch its own app.
  • These copycats often wrap a weaker, free version of OpenAI's model and charge users, though the App Store is beginning to clamp down on this practice.
• User Experience Blind Spot: Average consumers often cannot distinguish between different model versions (e.g., GPT-3.5 vs. GPT-4), missing the loss of capabilities like internet access or vision in copycat apps.

Emerging Web Trends & Company Movers

• Productivity Category: New entrants like Gamma (AI presentation builder) and ElevenLabs have made significant jumps on the web list, signaling a shift toward workspace integration.
• Companion Apps Growth: The number of companion apps on the list doubled from two in the first iteration to eight in the most recent.
  • Many of these utilize uncensored or loosely filtered models (e.g., via Llama) to offer greater conversational freedom compared to restrictive enterprise models like ChatGPT.
  • Gamification: Companies like Talkie integrate "trading card" mechanics where users unlock assets by spending time with companions, which are then tradable in in-app marketplaces.
  • User Intensity: Character AI reports an average user session count of ~300 per month (approx. 10 daily), with session durations averaging over five minutes, rivaling iMessage and Instagram usage.
  • Vertical Limitations: While consumer/casual companion apps go viral quickly, professional use cases (therapy, coaching) are growing but lack the same immediate viral volume.

Globalization of AI Talent Hubs

• AI development is shifting from a Silicon Valley-centric model to a global distribution of talent hubs emerging from universities and research labs worldwide.
• Key International Examples:
  • Mistral (France).
  • PhotoRoom (France), which ranks on both web and mobile lists.
• Barriers to Entry: The opening of models via API or open-source distribution allows solo developers in regions like China or Russia to build competitive products without massive capital raises.

Predictive Signals & Community Dynamics

• Discord as a Leading Indicator: Discord traffic serves as a proxy for future product success; companies like Midjourney, Pika, and Suno launched web apps immediately after establishing large Discord bases.
• Midjourney's Dominance: Midjourney accounts for nearly 1% of all Discord invite traffic, maintaining high engagement over 18+ months.
• Data Granularity: Analytics allow tracking of specific subdomains (e.g., discord.com/invite/[string]) to identify high-growth AI communities before they launch public web applications.

Forward-Looking Outlook (Next 6 Months)

• Audio as Primary Input: Expect a surge in mobile products leveraging audio capture to transform spoken ramblings into structured content (emails, notes) rather than simple transcription.
• Accessibility & Demographics: Voice AI is becoming a critical interface for older demographics (for companionship and information) and non-native speakers (for language practice and translation).
• Ecosystem Maturation: The industry is predicted to follow a trajectory similar to the early web, where infrastructure (blogs) precedes distribution channels (podcast apps) by roughly a decade.

Disclaimer

• Content is for informational purposes only and does not constitute legal, business, tax, or investment advice.
• A16Z and its affiliates may maintain investments in the companies discussed.