newsfilter.io
Interview

The Top 100 Most Used AI Apps in 2025

a16zOlivia Moore, Justine Moore
Watch on YouTube →

  • Methodology & Scope:

    • This is the fifth iteration of the A16Z Consumer AI Top 100 list, released every six months to track actual consumer usage rather than revenue.
    • Data is aggregated from SimilarWeb (monthly global visits for websites) and Sensor Tower (monthly active users for mobile apps).
    • The list selects the top 50 AI-native companies from each source, including both free and paid usage to measure genuine consumer imagination.
    • The ecosystem is stabilizing; the number of new entrants dropped from 17 on the web list six months ago to 11 in this edition.

  • Major Market Shifts & New Entrants:

    • Vibe Coding: A new dominant category, with Replit and Lovable entering the top 50 and Bolt landing just below the cutoff.
    • Companionship: Continues to dominate the landscape; three new entrants (Juicy Chat, Joy, R Dream) joined established names like Character AI, Spicy Chat, and Candy AI.
    • Grok: Debuted at #4 on the web list, driven by the release of Grok 4 and its companions and image/video features.
    • Meta AI: Made a visible entry on the web list, though it has not yet cracked the top 50 mobile list.
    • Mobile vs. Web Stability: The mobile list was historically unstable due to copycat "Chat" apps, but iOS policy crackdowns have reduced this noise, leading to more diverse new names.

  • Google's Strong Performance:

    • Gemini: Ranked #2 on the web list with roughly 10% of ChatGPT's traffic; on mobile, it reaches 50% of ChatGPT's traffic, driven primarily by Android users.
    • AI Studio: Surprisingly ranked in the top 10 as a developer-facing sandbox for testing Google models.
    • Notebook LM: Ranked #13, maintaining consistent flat or increasing traffic since its viral launch in September 2024.
    • Google Labs: Ranked #39; traffic spiked 15% in the month of VO3 (the video model) release.
    • Google Labs Portfolio: Includes experimental tools like VO3 (video), Doppel (outfit try-on), Portrait (coaching), Wisk (sandbox), and Project Mariner (agentic browser).

  • The Rise of Chinese AI Companies:

    • Domestic Usage: Products banned in the US for China (e.g., due to data regulations) dominate the top 20 web rankings, including Quark (Alibaba), Daobao (ByteDance), and Kimmy (Moonshot AI).
    • Global Exports: Chinese startups like DeepSeek, Hailuo, Kling, and C-Art are gaining traction internationally, particularly in image and video generation where they compete with or beat US models (e.g., ByteDance's C-Dance).
    • Distribution Channels: Many Chinese models are accessed via US-hosted aggregators (e.g., CREA, Hedra) or developer platforms (e.g., Fal, Replicate) rather than direct consumer apps.
    • Hybrid Growth: Manus achieved a #1 ranking in Brazil and #2 in the US, with a reported $90 million annualized run rate.

  • Productivity & Vibe Coding Economics:

    • Revenue Retention: Vibe coding platforms show exceptional retention, with nearly 100% revenue retention over the first three months, suggesting strong enterprise or "prosumer" adoption.
    • User Behavior: Traffic to the "makers" (people using the tools) significantly exceeds traffic to the generated "outputs," indicating users are building personal software or using custom domains.
    • Future Mobile Potential: No vibe coding apps have yet ranked in the mobile top 50, displacing the heavily dominant companionship apps.

  • Consumer AI "All-Stars":

    • The List: 14 companies have appeared on every one of the five lists: ChatGPT, Perplexity, Poe, Character AI, Midjourney, Photobox, Leonardo, Cutout Pro, Veo, Eleven Labs, Quillbot, Gamma, Hugging Face, and Civit AI.
    • Model Aggregation: Over half of the "All-Stars" do not own proprietary models but succeed through superior UI, workflow integration, and model aggregation (e.g., Hugging Face, Civit AI).
    • Network Effects: Evidence of non-data network effects is emerging in communities (Hugging Face, Civit AI) and asset libraries (Eleven Labs voice library).
    • Bottom-Up Adoption: Tools like Gamma and Eleven Labs are seeing organic enterprise spread, where individual users drive team adoption without traditional sales cycles.

  • Predictions & Future Trends:

    • Verticalization: Users are increasingly adopting specialized tools for specific tasks (e.g., Gemini for research, Claude for writing) rather than relying on a single generalist app.
    • Reliability Explosion: As foundation models (e.g., Grok 4, Claude 3.5, GPT-5) improve in logic and math, productivity tools (financial modeling, spreadsheets) will see a surge in usage and revenue as hallucinations decrease.
    • Emerging Categories: Next editions are expected to feature more AI-native EdTech, personal finance apps, and structured health/wellness coaching platforms.
    • Social AI: A dedicated AI-native social platform is predicted, with Grok potentially filling this role by combining chat and social feed capabilities.
    • Health Sector: Improved model accuracy is expected to drive a wave of reliable consumer health products in the near future.