The Top 100 Consumer AI Apps: Who’s Actually Paying?
a16z PodcastFull Title
The Top 100 Consumer AI Apps: Who’s Actually Paying?
Summary
This episode of the a16z Podcast discusses the seventh edition of their Top 100 Consumer AI Apps Report, introducing consumer spending data for the first time.
The report reveals that while a large portion of Americans are using AI, a smaller fraction are paying users, with developers and creators being the biggest spenders, and highlights the rise of personal agents and potential new business models for AI.
Key Points
- Consumer AI adoption is widespread, with about half of Americans using AI, but paying users are a small, concentrated group, with the top 1% spending significantly on AI tools.
- The AI app landscape is stabilizing in terms of traffic, with fewer new entrants compared to previous years, but revenue data shows a different picture with many new successful apps not prominent in traffic.
- Personal agents are emerging as a significant new trend, with companies like OpenAI (through Dots) and others developing more consumer-friendly and safe AI tools.
- High spending on AI tools by power users is often for development, productivity, and creative purposes, indicating AI is used as a tool for creation and enhancement rather than just task completion for the mainstream.
- The cost to serve AI users is a major factor influencing business models, with high inference costs currently leading to a greater reliance on subscriptions, though other models like ads and transaction fees are expected to emerge.
- OpenAI's rapid success with advertising, reaching a billion-dollar annual run rate, suggests a strong market potential for ads within AI platforms, especially as they gather more user data.
- The major AI labs (OpenAI, Anthropic, Gemini) are evolving differently, with ChatGPT maintaining dominance, Claude surpassing Gemini in paid subscribers, and Anthropic focusing on a subscription-only model.
- While big labs excel at core AI models, specialized AI products in areas like audio generation (11 Labs, Suno) and niche creative tools are finding success by focusing on specific user needs and interfaces.
- Many consumer AI categories remain underserved, particularly social AI, dating, recruiting, shopping, and entertainment, presenting significant opportunities for new startups.
- The future of AI consumer products will likely involve a strong software layer and specialized interfaces that leverage AI models, rather than just being a direct wrapper around the models themselves, creating compounding value for users.
Conclusion
The consumer AI landscape is evolving rapidly, with a clear trend towards power users driving revenue and the emergence of personal agents as a key area of innovation.
While subscription models are currently dominant, the future will likely see a diversification of business models, including advertising and transaction fees, as AI costs decrease and new use cases emerge.
Significant opportunities remain for startups in underdeveloped categories like social, dating, entertainment, and shopping, as well as in building specialized software layers that leverage AI models to create unique user experiences.
Discussion Topics
- What are the most compelling use cases for AI agents that could drive widespread consumer adoption beyond current productivity gains?
- How can AI companies balance the need for user data to improve services and enable targeted advertising with growing consumer privacy concerns?
- Which underserved consumer categories, such as dating, social networking, or entertainment, do you believe are most ripe for disruption by AI-native startups?
Key Terms
- Consumer AI Apps
- Software applications designed for individual consumers that utilize artificial intelligence technologies.
- Personal Agents
- AI-powered software designed to assist users with tasks, manage information, and automate actions based on user preferences and context.
- Power Users
- Individuals who use technology or software extensively and often in advanced or specialized ways, frequently for professional or creative purposes.
- PLG (Product-Led Growth)
- A business strategy where a product itself is the primary driver of customer acquisition, retention, and expansion.
- Inference Costs
- The computational cost associated with running an AI model to produce an output (e.g., generating text, an image, or a response).
- COGS (Cost of Goods Sold)
- The direct costs attributable to the production of goods or services sold by a company.
- IP (Intellectual Property)
- Creations of the mind, such as inventions, literary and artistic works, designs, and symbols, names and images used in commerce.
- Diffusion Model
- A type of generative model in machine learning used for creating new data, particularly images, by gradually adding and then reversing a diffusion process.
- AGI (Artificial General Intelligence)
- A hypothetical type of artificial intelligence that possesses the ability to understand, learn, and apply knowledge across a wide range of tasks at a human level.
- Network Effects
- A phenomenon where a product or service becomes more valuable as more people use it.
- Marketplace
- A platform that facilitates transactions between buyers and sellers.
- SaaS (Software as a Service)
- A software licensing and delivery model where software is licensed on a subscription basis and is centrally hosted.
- OpenAI
- A research laboratory focused on artificial intelligence, known for developing models like GPT-3, GPT-4, and DALL-E.
- Anthropic
- An AI safety and research company founded by former OpenAI employees, known for developing the Claude family of AI models.
- Gemini
- A large language model developed by Google, designed to be multimodal and capable of understanding and operating across different types of information.
- Monetization
- The process of converting something into money.
- Consumer Investors
- Investors who focus on companies and products that cater to individual consumers.
- Incumbents
- Existing companies or organizations that hold a dominant position in a particular market.
- Cannibalize
- To destroy or reduce the value of one's own products or services by introducing a new one.
- Inflection Points
- Critical moments where a significant change or development occurs.
- Software Layer
- The application or user-facing part of a technology stack, distinct from the underlying hardware or models.
- Hardware
- The physical components of a computer system or device.
Timeline
Half of Americans now report using AI, but only a small fraction are actually paying for it, and the people who do are spending a lot.
This episode is dedicated to discussing the seventh edition of the "Top 100 Consumer AI Apps Report," focusing on payment data.
On a traffic basis, many AI products seem to be settling, with fewer new products appearing.
The most interesting big new trend is in personal agents, with companies like OpenClaw pioneers and newer safe consumer-facing agents emerging.
AI has shifted from a productivity enhancement tool to something that can actually get things done for users, with OpenClaw being a significant pioneer.
While there aren't many new AI entrants, existing ones are getting larger, with a focus on consumer assistance like Muse and Instinct.
Muse has seen significant downloads and active users, particularly within its community, but its overall appeal to the average person is less than established social platforms like Threads.
The cost to serve AI users is high, which influences distribution strategies, and Muse does not need to push as hard as social media platforms due to its personal nature.
Network effects and marketplaces are crucial for AI agents to stick around long-term.
Personal AI assistant products haven't yet unlocked person-to-person network effects due to privacy concerns and the highly personal nature of the data they handle.
Privacy, safety, and security are increasingly important for consumer AI adoption, as users need to trust the tools they give intimate access to.
The resolution of how AI agents interact with platforms (endogenous) and consumer comfort levels (exogenous) are key questions for AI adoption.
While platform capabilities are improving, consumer willingness to use AI is still developing, with high costs to serve users being a concern.
The number one use case for agents, even consumer-focused ones, is still coding and technical automation, contributing to high service costs.
The flourishing of AI ideas outside of coding and productivity is limited because the personal nature of AI makes sharing and learning from others difficult.
Approximately half of Americans use AI, but only about 4.5% pay for AI products, with spending heavily concentrated among a small percentage of users.
The top 1% of paying AI users spend an average of $903 per month, highlighting the power user dynamic.
The products most paid users spend on are primarily developer tools, productivity tools, and creative tools, focusing on building and making.
High spending on AI tools by power users enables a new generation of makers to create art, videos, and content more efficiently.
Developers and coders have a mindset geared towards leverage and building products that maximize time, a worldview that is still reflected in a smaller user group.
Many consumer categories are still wide open for AI, especially those focused on how people spend their time rather than just saving time.
The next wave of AI users might be those who want to create videos or other content but lack the skills for professional tools, with AI enabling them to bring ideas to life.
The current model of direct subscriptions for consumer AI revenue is unsustainable, and other business models like ads and transaction fees will become more important.
OpenAI has achieved a $1 billion annual run rate on advertising, demonstrating significant market traction.
OpenAI's advertising success is attributed to its large user base and the potential for more targeted ads due to user data.
The implementation of ads in AI needs to be done carefully to maintain user trust, especially given the personal nature of AI interactions.
AI is increasingly being used for real advice, shopping, and discovery, creating opportunities for ads in a "discovery mode."
Verticalized consumer AI products, like OpenEvidence for doctors, can be successful with advertising due to targeted, high-value audiences.
ChatGPT remains dominant in consumer AI usage and revenue, while Claude has overtaken Gemini in paid subscribers among US users.
Anthropic (Claude) is more aggressive with subscriptions and has a higher percentage of users on its premium plan, emphasizing a no-ads policy.
Big AI labs are good at creative tools, but specialized companies are excelling in specific modalities like audio generation (11 Labs, Suno).
On images, OpenAI and Google are strong, with Midjourney still performing well in revenue rankings among power users who value its aesthetic sense.
Video generation is a complex frontier with Chinese companies having an advantage due to data training.
Creative tools that are agent-friendly and accessible will likely see strong tailwinds.
The trend of consumer companies rapidly expanding into enterprise offerings is driven by successful PLG strategies and the work-productivity nature of AI tools.
Incumbents are often unwilling or unable to cannibalize their existing interfaces, creating opportunities for startups to reinvent products for the AI age.
The improvement in AI models has led to better products and increased demand, with the software layer and user experience being key differentiators.
While big labs like OpenAI and Anthropic will remain successful, startups have opportunities to innovate and capture market share with specialized products.
The value is shifting back to the software layer and user experience, with AI models being a component rather than the sole asset.
Compounding personal and multiplayer value in AI products creates a "lock-in" effect, making it difficult for users to migrate.
Many categories remain open for AI innovation, particularly those requiring multiplayer products and network effects.
Most pure consumers are using AI as a replacement for search engines, or for specific use cases like education and workplace tasks.
Categories like dating, recruiting, social AI, shopping, and entertainment are wide open for AI-native startups.
While AI agents excel at conversational tasks, shopping and entertainment require visual exploration and customization, areas ripe for AI innovation.
The value in consumer AI will ultimately reside in the software layer, the process, and the user experience, as models continue to improve.
The focus of consumer AI has largely been on "saving time," but there's potential for AI to enable "spending time" more creatively and enjoyably.
Episode Details
- Podcast
- a16z Podcast
- Episode
- The Top 100 Consumer AI Apps: Who’s Actually Paying?
- Official Link
- https://a16z.com/podcasts/a16z-podcast/
- Published
- October 5, 2026