Are AI Girlfriends a Thing? (Yes, Here's the Scale of It)

How big the AI girlfriend market actually is

I maintain a live database of AI girlfriend platforms that I personally test and re-test, and as of our most recent full audit on July 6, 2026, it holds 129 distinct platforms. That's not a list padded with dead links or abandoned apps. Every one of them is a functioning product with its own pricing, its own feature set, and its own user base at the time of testing.

129 is a big enough number that no single person could plausibly build all of these as a side project. Behind that number sits real venture funding, real engineering teams, and real competition for the same audience, which is a much better indicator of market size than any single company's self-reported user count would be.

It's not one product, it's a whole product category

A real market doesn't just have a lot of competitors, it has meaningfully different types of products competing for different slices of demand. That's exactly what shows up across our 129 platforms. Some are pure text chat. Others lean almost entirely on AI-generated images. A smaller group has invested specifically in voice, and a newer group has pushed into AI-generated video.

That NSFW-versus-SFW split alone tells you this isn't a single homogenous product with minor cosmetic differences. It's a category with distinct sub-markets, each attracting a different type of platform and, almost certainly, a different type of user.

Scoring also shows this diversity isn't just cosmetic. Across the five categories we test, chat quality averages 3.26 out of 5, pricing averages 3.30, image generation sits at 2.12, voice at 1.81, and customer support at 2.21. Those aren't five versions of the same number, they're five genuinely different competitive fronts, which is exactly what you'd expect from a market with real specialization rather than 129 near-identical clones of the same product.

The pricing tells its own story

Markets that aren't real don't develop consistent pricing structures, because there isn't enough sustained demand to support one. This one clearly has. Out of our 129 tracked platforms, 16 are genuinely free, 56 sit in a budget tier, 55 in mid-range, and only 2 have pushed into premium pricing. The average starting price across the 85 platforms with a clearly parseable price comes out to $11.85 a month.

That clustering in the budget-to-mid-range band is itself informative. It means the market has more or less converged on what people are actually willing to pay for this kind of product, the same way any maturing consumer software category eventually settles into a workable price band instead of guessing wildly. If you want to skip the guesswork entirely, our own best AI girlfriend rankings score every platform we track on pricing alongside chat quality, memory, voice, and images.

The churn: a young, fast-consolidating market

Real markets also churn, especially young ones, and this one churns hard. When I re-audited every platform in our database this year, at least 23 of them, about 18%, had either gone completely dark, been quietly sold to a new owner, or rebranded beyond recognition under the same old name. That's close to 1 in 5 platforms changing meaningfully in a single year.

That's not a sign the category is failing. It's actually the opposite: it's what active consolidation looks like in a fast-growing, still-forming market, where weaker or undifferentiated products get absorbed, abandoned, or replaced while the overall pool of platforms keeps refilling with new entrants. Categories that aren't real don't have enough activity to produce this kind of churn in the first place.

Why the category doesn't always feel that big from the outside

Part of why people underestimate the scale of this market is that it's built around private, one-on-one use rather than public sharing. Nobody posts a screenshot of their AI girlfriend conversation the way they'd post a photo from a night out. That privacy is a feature of the category, not a flaw, but it also means the actual footprint of the market is much bigger than its visible footprint online.

The other reason is fragmentation. With 129 competing platforms rather than two or three dominant ones, no single brand accumulates the kind of cultural visibility that would make the category feel obviously large the way, say, a single dominant social app would. The size is real, it's just distributed across a lot of separate products instead of concentrated in a couple of household names.

Discretion around payment reinforces the same pattern. 19% of the platforms we track accept cryptocurrency payments specifically, a genuinely high figure for a mainstream consumer software category. That's a direct signal of how much this industry has had to build around users who don't want a subscription to an AI girlfriend app showing up by name on a shared bank statement. A market that size doesn't need to build discreet payment options unless there's real, sustained demand for them.

So who's actually behind these numbers?

Scale and product diversity tell you the category is real and commercially significant, but they don't directly tell you who's on the other end of these conversations or how often. That's a distinct question worth answering on its own terms, and it's exactly what I dig into in a separate piece on who's actually using AI girlfriend apps, using free tier adoption, pricing behavior, and category diversity as the honest proxies we actually have.

Is this a fad, or is it structural?

Fads tend to concentrate around one product, spike fast, and collapse just as fast once the novelty wears off. What we're tracking instead is 129 independently built platforms, distinct sub-categories serving different needs, converged pricing, and a churn pattern that reflects consolidation rather than collapse. That's the signature of a maturing product category, not a short-lived trend.

The quality is genuinely uneven, our own average score across all 129 platforms sits at just 2.5 out of 5, but uneven quality within a category is completely different from the category itself not being real. Plenty of large, legitimate markets have a long tail of mediocre products alongside a handful of genuinely good ones, and this one is no exception.

It's also worth noting that the 2.5 average holds steady whether you look at NSFW-allowing platforms or SFW-only ones. Both groups land at exactly 2.5 out of 5 in our testing. That consistency across such different product types is itself a small piece of evidence that we're measuring one coherent market with shared underlying dynamics, not two unrelated things that happen to share a label.

How we track and verify all 129 platforms

Every platform in our database gets tested directly, not just researched from a distance. I use a real subscription on each one, run actual conversations, test whatever voice, image, or video features exist, and score each platform across five categories on a consistent scale. That's also how we catch the churn we mentioned earlier, since re-testing an existing platform is exactly what surfaces a dead link, a quiet sale, or a rebrand.

You can read the full breakdown of exactly how this works on our testing methodology page, and more about the process behind it on my author page. If you're still getting your footing on what an AI girlfriend actually is before diving into the scale of the market, start with our complete definition.

Further reading

• Do People Really Use AI Girlfriends?
• What Is an AI Girlfriend? A Complete Definition
• The best AI girlfriend apps, ranked