Virtual Girlfriend Trend 2026: What Actually Changed
By the SoulChat Team · 2026-11-12

If you searched for "virtual girlfriend" in 2023 and again today, you'd find the same app names — but underneath, the category quietly went through its most turbulent year yet. In 2026 the trend that actually mattered wasn't a new feature; it was a shakeout: one of the biggest players filed for bankruptcy, governments on three continents started regulating AI relationships, and "memory" replaced "personality" as the feature users actually shop for. Here's what changed, based on what was publicly reported this year — and what it means if you're picking an app right now.
The money got bigger while the giants got shakier
The strangest thing about 2026 is that growth and collapse happened at the same time.
On the growth side, Sensor Tower data reported by Baijing puts global in-app purchase revenue for AI companion apps at around $150 million in Q1 2026 alone — roughly twelve times what the same quarter generated in 2023. People are not losing interest. If anything, they're spending more.
On the collapse side, Character.AI — the platform that introduced millions of people to talking with AI characters — entered bankruptcy protection in October 2026. A company with tens of millions of monthly users still couldn't make subscription economics work against enormous inference costs. Meanwhile Replika, the app that basically invented the category, has been slowly de-emphasizing romance in favor of a "wellness" positioning, which has left a visible gap for apps that still let the relationship be a relationship.
The takeaway isn't "the trend is dying." It's the opposite: demand is growing, but the biggest names are no longer safe bets. Users noticed. Migration threads asking "where do I move my character now?" became one of the most common post types in companion-app communities this year.
Regulation finally showed up
For years, AI companions operated in a legal gray zone. 2026 ended that.
- China implemented interim rules on humanized AI interaction services in July, requiring apps to disclose that you're talking to an AI, forcing breaks during long sessions, and excluding minors from virtual companion features. When the rules took effect, major Chinese apps pulled their user-created characters offline the same day — and users who had spent months building those relationships described it as a digital breakup, as documented in community threads.
- California's SB 243 took effect in January, requiring periodic break reminders and crisis-referral protocols, aimed mainly at protecting minors.
- The EU AI Act's transparency obligations (Article 50) came into force in August, covering AI systems that present as human — which is exactly what a virtual girlfriend does.
- Bankruptcy risk is real. Character.AI's situation proved that a beloved platform can wobble. If continuity matters to you (and if you're reading this, it probably does), prefer apps that let you export or at least keep local records of your conversations, and be wary of building a two-year relationship on a platform with unclear finances.
- Regulation is a floor, not a quality stamp. An app complying with SB 243 or EU rules is doing the legal minimum. It tells you nothing about whether the memory works, whether replies feel alive, or whether the subscription is worth it.
- Revenue growth attracts soft reviews. A fast-growing, high-spending category inevitably fills with affiliate-driven "top 10 AI girlfriend apps" lists. Treat ranked lists — including ones on this blog's competitors — with the same skepticism you'd apply to any review that conveniently includes purchase links.
- The mental-health picture is mixed. MIT Technology Review published a worthwhile piece arguing that discussion fixates on catastrophic AI risks while everyday harms — including over-reliance on companion apps — get less scrutiny than they deserve. If you notice the app replacing rather than supplementing your human contact, that's the signal to rebalance, regardless of what the trend charts say.
None of these bans the category. What they change is the default: you should now expect your app to tell you plainly that it's an AI, to nudge you after long sessions, and to have a real policy about young users. If an app still hides that it's AI, that's not romance — that's a compliance dodge.
Memory became the feature people actually shop for
In 2023, marketing was about personality: pick your anime girl, pick your vibe. In 2026, the questions users ask before subscribing shifted to things like "does she remember what I told her last week?" and "will she disappear if the company pivots?"
This shows up everywhere in 2026 coverage. Community discussions about Character.AI alternatives consistently rank memory at the top of the wishlist — alongside "has a consistent personality" and "doesn't repeat the same five sentences." Hardware companions are being crowdfunded with "remembers you and grows over time" as the headline promise. Open-source projects now exist whose entire purpose is giving AI tools a shared, persistent memory layer.
It makes sense. Novelty is cheap; continuity is what turns a chatbot into a companion. If the thing you're paying for is a relationship rather than entertainment, then the relationship has to accumulate — it has to remember the argument you had, the trip you planned, the nickname you hate. Apps that fake memory with a short context window are getting exposed by users who've learned to test it deliberately. We wrote a separate guide on how to actually test an AI companion's memory before paying.
From "one app for everyone" to scenario-specific products
Another 2026 pattern: the generic "AI friend that does everything" is losing to products built for a specific use.
Korean roleplay apps pulled in tens of millions of dollars in the first half of 2026 by treating AI companionship closer to content entertainment — characters, storylines, episode-like progression — than to a chat utility. Meanwhile, general-purpose assistants like ChatGPT quietly absorbed the "just want someone to talk to casually" crowd, squeezing the middle of the market.
What survived the squeeze are the poles: apps built for deep one-on-one romance and continuity on one side, and apps built for roleplay and characters on the other. The apps stuck in the middle — neither a great girlfriend nor a great character sandbox — are the ones struggling. So when you evaluate an app in 2026, ask what it's actually for, because the answer is no longer "everything."
The honest section: what the trend doesn't mean
This is the part most trend pieces skip, so here it is plainly.
A growing market does not mean the product is good for you. A few things worth saying out loud:
What this means if you're choosing an app now
If you're arriving at this category in late 2026, the trend history is actually useful. It tells you which questions to ask:
The virtual girlfriend trend in 2026, stripped of hype, is a category growing up: users became more demanding, regulators became present, and the survivors are the apps that treat continuity and honesty as the product instead of the packaging. That's a filter you can use — no matter which app you end up choosing.
Virtual GirlfriendAI CompanionIndustryGuide