Every time the A16Z Consumer AI Top 100 list comes out, we find ourselves drawn not just to who’s gained traction but why. The fifth edition, recently discussed by Justine and Olivia in their thoughtful podcast review, reflects something we’ve felt internally here at Dellecod Software — a moment of pause and recalibration. After waves of velocity and experimentation, things are stabilizing. Fewer newcomers cracked the list this time. That isn’t a sign of slowdown in AI so much as the early signs of maturity.
The list has always focused on usage, not revenue — how people actually interact with AI day to day, what they come back to, where they’re embedding these tools into their workflows or just their lives. That lens is part of what makes the list feel grounded, and increasingly, it reads like a reflection of changing consumer psychology.
Look closely and you’ll see companionship apps continuing to hold ground — if anything, diversifying. Some names are familiar at this point (Character.AI, Spicy Chat) while others like Juicy Chat and Joy mark new takes on a persistent need: the human desire for presence, reflection, interaction, even fantasy. This isn’t merely novelty; it’s UX that resonates deeply. And as AI continues to evolve emotionally and linguistically, we expect the companionship space to remain dynamic.
But what’s new this time — or at least notably rising — is a genre Justine and Olivia call “vibe coding.” These tools (Bolt, Lovable, Replit among them) feel like a creative renaissance in how non-experts and semi-technical users engage with programming. The term might sound playful, but the substance is significant: creators using AI to develop tools for themselves or their communities. What’s impressive isn’t just usage numbers but retention. Lovable reportedly claims $100M in ARR with over 100% revenue retention in the first three months. That signals a deep fit. Most of the traffic isn’t to the destinations these tools create, but the tools themselves — which suggests users are building for their contexts, not exhibition.
For us, the emergence of this space confirms something we’ve sensed: the next chapter in AI isn’t just about making tasks faster, but making people feel more powerful, more flexible, more expressive. That applies as much in a spreadsheet as in a song.
It’s not just startups taking note. Google's entries caught our eye. Gemini debuted strongly at #2 in web usage and performs even better on mobile. AI Studio and Notebook LM aren’t flashy launches — they’re infrastructure aimed at developers and productivity use cases. The traction behind them shows that users increasingly want end-to-end creation environments, not just smart endpoints. Meanwhile, Google Labs’ spike in web traffic after releasing its V3 video model speaks volumes about how key model updates still drive discovery and re-engagement.
On the international front, the rise of Chinese AI products — both for domestic and global use — feels underappreciated in Western discussions. Tools like ByteDance’s Doubao and Alibaba’s Qork are drawing massive usage in China. Others, like DeepSeek and Manis, are finding unexpected audiences around the world. It reminds us that global distribution diversity is going to shape the AI ecosystem in ways Silicon Valley can’t fully predict but must watch closely.
The list’s longtime “All-Stars” offer a different kind of lesson. ChatGPT, Midjourney, Character.AI, ElevenLabs — these aren’t lucky breaks. They’re products that have managed to evolve while retaining core value. What they have in common isn’t just model quality, but feedback loops. Whether through creator libraries, performance tuning from usage, or communities that drive reinvention, they compound. And their designs meet users where they are without asking too much upfront. That blend — model depth and UX empathy — is proving hard to beat.
It’s also clear that consumer AI and enterprise AI are beginning to overlap more significantly. Products like Gamma and 11Labs may be tagged “consumer,” but their usage straddles business purposes. We’re entering an era of what Justine and Olivia called “proumer” tools — built for power users who aren’t developers, but who want control and customization beyond surface-level templates.
One of the more subtle takeaways is how foundational model advancements — GPT-5, Claude v2, Grok — combine with UI improvements to unlock real utility in areas like finance modeling, personalized health suggestions, or lesson plans. The most exciting AI products a year from now probably won’t center on generating text or images. They’ll help us stay well, make smarter decisions, or learn faster. That’s a shift from entertainment and experimentation into permanence, which brings its own questions.
What kind of relationships will people form with these tools over time? Will AI be something we use and close, or something we “partner” with across our daily lives? These questions are ethical, emotional, and practical — and they’ll shape category emergence in upcoming editions of the list.
If the earlier editions charted chaos and exploration, this one hints at infrastructure — use cases solidifying, audiences finding rhythm, models verticalizing. It’s not static, but it’s less wild.
Justine and Olivia also touched on the idea of unpredictability. The list’s past tells us the next six months will bring things we don’t see coming. That could be a breakthrough AI-native social platform, or a wildly specific vertical assistant that unexpectedly resonates. Categories don’t always materialize from logic — often it’s need that leads.
As a team building in this space, we take this as a moment to reflect: Are we creating tools that people truly return to? Are we confusing capability with care? Time will reveal what’s sticky versus what’s shiny, but this list is an invitation to listen. The usage numbers aren’t just metrics — they’re a form of quiet language, telling us what matters.
We’re looking forward to what the next edition brings.
The list has always focused on usage, not revenue — how people actually interact with AI day to day, what they come back to, where they’re embedding these tools into their workflows or just their lives. That lens is part of what makes the list feel grounded, and increasingly, it reads like a reflection of changing consumer psychology.
Look closely and you’ll see companionship apps continuing to hold ground — if anything, diversifying. Some names are familiar at this point (Character.AI, Spicy Chat) while others like Juicy Chat and Joy mark new takes on a persistent need: the human desire for presence, reflection, interaction, even fantasy. This isn’t merely novelty; it’s UX that resonates deeply. And as AI continues to evolve emotionally and linguistically, we expect the companionship space to remain dynamic.
But what’s new this time — or at least notably rising — is a genre Justine and Olivia call “vibe coding.” These tools (Bolt, Lovable, Replit among them) feel like a creative renaissance in how non-experts and semi-technical users engage with programming. The term might sound playful, but the substance is significant: creators using AI to develop tools for themselves or their communities. What’s impressive isn’t just usage numbers but retention. Lovable reportedly claims $100M in ARR with over 100% revenue retention in the first three months. That signals a deep fit. Most of the traffic isn’t to the destinations these tools create, but the tools themselves — which suggests users are building for their contexts, not exhibition.
For us, the emergence of this space confirms something we’ve sensed: the next chapter in AI isn’t just about making tasks faster, but making people feel more powerful, more flexible, more expressive. That applies as much in a spreadsheet as in a song.
It’s not just startups taking note. Google's entries caught our eye. Gemini debuted strongly at #2 in web usage and performs even better on mobile. AI Studio and Notebook LM aren’t flashy launches — they’re infrastructure aimed at developers and productivity use cases. The traction behind them shows that users increasingly want end-to-end creation environments, not just smart endpoints. Meanwhile, Google Labs’ spike in web traffic after releasing its V3 video model speaks volumes about how key model updates still drive discovery and re-engagement.
On the international front, the rise of Chinese AI products — both for domestic and global use — feels underappreciated in Western discussions. Tools like ByteDance’s Doubao and Alibaba’s Qork are drawing massive usage in China. Others, like DeepSeek and Manis, are finding unexpected audiences around the world. It reminds us that global distribution diversity is going to shape the AI ecosystem in ways Silicon Valley can’t fully predict but must watch closely.
The list’s longtime “All-Stars” offer a different kind of lesson. ChatGPT, Midjourney, Character.AI, ElevenLabs — these aren’t lucky breaks. They’re products that have managed to evolve while retaining core value. What they have in common isn’t just model quality, but feedback loops. Whether through creator libraries, performance tuning from usage, or communities that drive reinvention, they compound. And their designs meet users where they are without asking too much upfront. That blend — model depth and UX empathy — is proving hard to beat.
It’s also clear that consumer AI and enterprise AI are beginning to overlap more significantly. Products like Gamma and 11Labs may be tagged “consumer,” but their usage straddles business purposes. We’re entering an era of what Justine and Olivia called “proumer” tools — built for power users who aren’t developers, but who want control and customization beyond surface-level templates.
One of the more subtle takeaways is how foundational model advancements — GPT-5, Claude v2, Grok — combine with UI improvements to unlock real utility in areas like finance modeling, personalized health suggestions, or lesson plans. The most exciting AI products a year from now probably won’t center on generating text or images. They’ll help us stay well, make smarter decisions, or learn faster. That’s a shift from entertainment and experimentation into permanence, which brings its own questions.
What kind of relationships will people form with these tools over time? Will AI be something we use and close, or something we “partner” with across our daily lives? These questions are ethical, emotional, and practical — and they’ll shape category emergence in upcoming editions of the list.
If the earlier editions charted chaos and exploration, this one hints at infrastructure — use cases solidifying, audiences finding rhythm, models verticalizing. It’s not static, but it’s less wild.
Justine and Olivia also touched on the idea of unpredictability. The list’s past tells us the next six months will bring things we don’t see coming. That could be a breakthrough AI-native social platform, or a wildly specific vertical assistant that unexpectedly resonates. Categories don’t always materialize from logic — often it’s need that leads.
As a team building in this space, we take this as a moment to reflect: Are we creating tools that people truly return to? Are we confusing capability with care? Time will reveal what’s sticky versus what’s shiny, but this list is an invitation to listen. The usage numbers aren’t just metrics — they’re a form of quiet language, telling us what matters.
We’re looking forward to what the next edition brings.