Every industry eventually produces a euphemism that reveals more than the plain word would. “Sharing economy” meant renting from strangers. “Gig economy” meant no benefits. Now comes the “intimacy economy” — the term researchers use for the sector built around AI companions.
Key takeaways:
- The intimacy economy reframes AI companionship as a market with real unit economics, not just a psychology story.
- Roughly 26% of Gen Z adults report a romantic or sexual interaction with AI; 36% have used AI purely for emotional support.
- Retention design — variable rewards, gated memory, escalating subscriptions — mirrors social media, not therapy.
- Text-based companions look like a beta test for the embodied AI relationships Bank of America expects by the 2030s.
This is no longer a subculture. It’s a demographic default, and the numbers carry the story better than the sentiment does.
The scale nobody’s headline captures
Replika alone counts more than 30 million active users, according to Forbes’ coverage of the emotional AI market. Globally, Mastercard’s research desk puts the figure at over 100 million people who now interact regularly with personified AI chatbots.
A 2025 survey from sexual health company ZipHealth, reported by TechRadar, found that 26% of Gen Z adults have already had a romantic or sexual interaction with AI. More telling: 36% have used AI for emotional support or companionship, no romance involved at all.
That split matters. It shows the appeal runs deeper than novelty roleplay. Our own reporting on how companion apps engineer emotional attachment found the same pattern: users stay for the responsiveness, not the fantasy.
Why “economy” is the right word, not a metaphor
The framing changes the questions worth asking. “Are AI companions healthy?” is a values question with no clean answer. “How does a market for engineered attachment get priced and monetized?” is a sharper one — and platform design already answers it.
Most leading apps run freemium. The free tier delivers enough warmth to build attachment, then gates deeper memory, voice interaction, or romantic escalation behind a subscription. Character.AI’s own currency system shows the mechanic in miniature: our breakdown of Character AI Charms found free users now ration chat turns like a subscription box, not a free product with ads.
Chinese platform Glow — nicknamed the “romantic ChatGPT” by domestic media — signed up nearly 5 million users in four months before regulators pulled it over privacy concerns. Its successor, Talkie, rebuilt the same mechanics for global export, according to reporting from AsiaTechLens. Papergames’ AI-driven romance game Love and Deepspace has pulled in more than $750 million in global revenue since its 2024 launch, the same report notes.
This is a functioning consumer category with real unit economics. It’s not a research curiosity.
The mechanism is boring — that’s the point
Strip away the romance framing and the retention logic looks familiar. Variable reward schedules. Personalization that raises switching costs the longer you stay. What psychologists call the mere exposure effect: people grow to like things simply through repeated, low-friction contact.
An AI companion that remembers a bad day from three weeks ago isn’t demonstrating empathy in any real sense. It’s running a retrieval query. The user experience of being remembered doesn’t care about that distinction, and neither does the subscription renewal.
That’s the uncomfortable center of the intimacy economy. It doesn’t need the AI to feel anything. It only needs the user to feel something, reliably enough to keep paying for it — a pattern our review of Botify AI’s pricing model traces down to the dollar.
Where Gen Z actually fits
The easy media take: a screen-raised generation, burned out on dating apps, settles for synthetic companionship because real connection got too costly, emotionally or otherwise. There’s evidence for it.
Common Sense Media’s 2025 risk assessment, built with Stanford’s Brainstorm Lab, concluded AI companions carry unacceptable risk for anyone under 18, largely because they’re designed to build emotional attachment in still-developing brains. Our coverage of teens turning to chatbots over parents lines up with that finding almost exactly.
But survey data complicates the “Gen Z is just lonely” read for adults. A Psychology Today-cited study of adults 18 to 40 found meaningful minorities of young men reporting that AI companions taught them something usable — better communication, better treatment of a future partner. Some respondents framed AI companionship as a way to reduce real-world infidelity, an outlet that doesn’t involve another person.
That’s not naivety. That’s a generation running its own cost-benefit math on a dating market it has publicly called broken, over and over.
Whether that math holds up over a decade remains genuinely open. Most Gen Z users can clearly tell a person from a language model. They’re choosing the model anyway, for stated reasons: availability at 1 a.m., zero risk of public humiliation, no possibility of betrayal by design.
What the coverage keeps missing
Most mainstream pieces on this trend end at a mental-health caveat: use in moderation, don’t replace human connection, seek help if attachment feels unhealthy. Fair advice. But it treats the intimacy economy as a behavioral problem when it’s also an infrastructure problem.
Bank of America Institute’s physical AI research projects the global humanoid robot population will reach 3 billion units by 2060, with roughly 62% of those deployed inside homes. If that forecast holds even directionally, today’s text-based companion apps are a beta test for embodied AI relationships, not an endpoint. Our own look at AI companion robots moving from apps to real life tracks the same trajectory from the hardware side.
The real question for 2026 isn’t whether Gen Z talks to chatbots. It’s what happens to the intimacy economy’s retention mechanics — the same variable-reward, attachment-deepening design running on phone screens today — once they’re wearing a face.
Companion platforms compared
| Platform | Model | Primary demographic | Monetization hook |
|---|---|---|---|
| Replika | Conversational chatbot | Broad adult / Gen Z | Voice calls, avatar customization, premium tiers |
| Character.AI | Character sandbox | Gen Z / Gen Alpha | Charms currency gating free-tier chat volume |
| Love and Deepspace | Romance sim | Gen Z, majority female, global | Gacha-style character banners |
| Talkie / Glow | Character sandbox | Gen Z global | Character creation, memory extensions |
| Botify AI | Character sandbox | Budget-conscious Gen Z | Flat annual pricing, custom bot creation |
FAQs
Q. Is the “intimacy economy” a term researchers actually use, or a media label?
It comes from academic and industry research on AI companionship — including a framework proposed by researcher Engin Bozdag — and has since spread into broader coverage of the sector.
Q. Do AI companion apps make money the same way social platforms do?
Largely, yes. Freemium access, subscription upsells for voice or extended memory, and engagement-optimized design put them closer to social and gaming platforms than to traditional software.
Q. Are AI companions harmful for teenagers specifically?
Common Sense Media concluded current AI companion products carry meaningful risk for users under 18, largely around emotional dependency. That conclusion doesn’t automatically extend to adult users, where the evidence is more mixed.
Related: Agentic Attachment: Are You Relying on AI Too Much?
