The Top 100 GenAI Products, Ranked and Explained

Market Timing and Maturity:

• Consumer AI adoption currently lags behind research and model development by 6 to 12 months; the sector remains in the early adopter/infrastructure-to-application transition phase.
• Historical assumptions regarding AI capabilities (e.g., inability to generate convincing video, resistance to AI music, or ChatGPT's permanent market dominance) are frequently invalidated months after being formed.
• The ecosystem is characterized by rapid iteration, with 17 new companies appearing on the web rankings in the most recent report and consistent entry of newcomers indicating the market is not yet stagnant.

Data Methodology and Reporting Structure:

• The "Gen AI 100" list is split into three distinct rankings: Top 50 on Web (via Similar Web), Top 50 on Mobile (via Sensor Tower monthly active users), and a new Top 50 on Mobile by Revenue.
• Revenue and traffic lists show only a 40% overlap, indicating that high-traffic apps do not necessarily correlate with high monetization.
• The newly introduced "BrinkList" captures the five web and five mobile products just below the cutoff, highlighting rapid shifts where incumbents like Runway and Otter have been temporarily displaced by new entrants like Deep Seek.
• 16 companies have appeared on every single web ranking across four report iterations, signaling that consumer AI businesses have achieved significant brand cementing despite the early market stage.

Key Pivotal Moments and Trends:

• AI Video: The market is "on the verge" of functional AI video, with three new video models (HiLuo, Kling, Sora) entering the top rankings; Sora's release and the performance of Chinese models (less copyright-sensitive) suggest a shift in capabilities.
• Vibe Coding: Tools like Bolt, Lovable, and Cursor have reached saturation among technical users, with an estimated 95% of YC companies building using these tools.
• Deep Seek Disruption: Launched at the end of January 2025, Deep Seek rose from zero to #2 on the web and #14 on mobile within one week, driven by free access to its reasoning model and real-time chain-of-thought visualization.
• Companion Apps: There is a high prevalence of companion/relationship apps on the list, including three in the top 10, often driven by NSFW or interactive fan fiction use cases, though multi-modality in this category remains underdeveloped.
• Chinese Model Influence: Chinese models are gaining traction due to fewer copyright constraints in training data and a higher density of researchers focusing on image/video modalities.

Revenue and Monetization Dynamics:

• High-revenue mobile apps (e.g., Speak, Otter, video editors) often operate with smaller user bases (1–2 million) compared to high-traffic consumer apps, suggesting a strategy of targeting "prosumers" who pay significantly higher per-user revenue.
• Many mobile apps utilize paid acquisition strategies that yield 1x–2x returns on user acquisition costs, allowing indie developers to sustain millions of users without Venture-style 10x growth expectations.
• Niche categories like plant identification apps generate significant revenue from small, highly dedicated user bases willing to pay high annual fees ($100/year).

Specific Product Performance and User Behavior:

• ChatGPT: Experienced a 2x increase in weekly active users (from 200M to 400M) over the last six months, driven by new models (O1, GPT-4o) and features like voice mode and deep research.
• User Stickiness: Deep Seek currently holds a 7% 30-day retention rate compared to ChatGPT's 9%, potentially skewed by geographic factors where Deep Seek is the default in regions where ChatGPT is blocked.
• Creation vs. Consumption: Traffic to creation platforms (e.g., Lovable.dev) significantly exceeds traffic to apps built on them, indicating untapped potential for viral products built by non-technical users.
• Mobile vs. Desktop: Mobile dominance is observed in voice-first products, avatar creation, and homework helpers, while desktop remains prevalent for complex coding and long-form video editing.

Future Outlook and Assumptions to Watch:

• AI Delegation: Contrary to the assumption that AI will handle transactional tasks while humans build relationships, AI is proving more effective at relationship building due to patience and nuance.
• AI as Manager: A shifting paradigm suggests AI may eventually delegate tasks to humans rather than the reverse, leveraging human creativity and joy in execution.
• Google VO2: Expected to launch within 3–6 months, potentially acting as the next-level benchmark in AI video generation.
• Disappearing Software: The cost of creation is dropping, enabling "disposable software" or short-shelf-life products that may only exist for minutes or days (e.g., personalized jokes or memes).
• Multi-modality Explosion: A projected surge in products leveraging new multimodal models (e.g., Gemini Flash, OpenAI Operator) for screen interaction and task execution is anticipated in the next 1–2 reporting cycles.

Investment and Strategic Takeaways:

• Success in consumer AI now depends on solving specific pain points or creating unique experiences rather than simply leveraging the newest model architecture.
• Product stability and user delight are more critical than technical novelty; teams are encouraged to use older models if they offer better utility or stability.
• The "chaos" of the app store is expected to increase as low-code/no-code tools allow non-technical founders to launch venture-scale products, necessitating new AI-driven discovery and management solutions.