AI Companion Market Size and Growth: What Our Own Data Suggests
We don't cite third-party market forecasts. Instead, here's what our own testing of 129 platforms suggests about growth: rising feature adoption, pricing experiments, and real churn.
Jordan Voss
AI Companion Researcher
December 10, 2025
Quick answer
We don't cite third-party market-research forecasts for AI companion market size, because we can't verify how those numbers are calculated. What we can share is our own directional data: we track 129 active AI girlfriend platforms, 22% now offer AI video generation (a feature category that barely existed when we started tracking), and at least 39 platforms added a free tier during a single re-audit pass, all signs of a market still actively expanding its feature set and its pricing experimentation. At the same time, at least 23 platforms, about 18%, went dark, were sold, or rebranded within a year, showing this growth comes with real consolidation, not just addition. That's the honest, own-data picture of where this market is headed.
People ask me for "the market size" of the AI girlfriend industry more than almost anything else, usually expecting a dollar figure and a growth percentage. I'm going to be straight with you: I don't have a verified, credible source for a specific market-size number, and I'm not going to invent one or repeat a figure from somewhere I can't stand behind. What I do have is a year-plus of direct, hands-on testing and re-testing of 129 platforms, and that data tells a real, useful story about growth, even without a dollar figure attached to it.
Why we're not citing a market-research forecast here
Plenty of sites will hand you a specific market-size number with a specific growth percentage attached. I'd encourage some skepticism toward those figures unless the source shows its actual methodology, since a lot of this category is private companies that don't publish revenue or user numbers, which makes a precise, verifiable market-size estimate genuinely difficult to produce honestly. Rather than repeat a number we can't independently verify, we're going to walk through what our own testing data actually shows, and be explicit about the difference between that and a dollar-denominated market forecast.
What our own data can actually tell us
Our dataset doesn't include revenue, user counts, or download figures for any platform, since most of that data isn't public and we're not going to guess at it. What we do track directly, because we test it ourselves, is platform count, feature adoption, pricing structure, and churn, all measured consistently across the same 129 platforms over time. Those are proxies for growth, not a market-size figure, but they're proxies built on data we've actually verified firsthand rather than numbers we're passing along from somewhere else.
Signal 1: the sheer number of active, competing platforms
We track 129 distinct AI girlfriend platforms today. That count alone tells you this isn't a niche corner of the internet with a handful of players, it's a genuinely crowded, competitive category. A market that couldn't sustain new entrants or ongoing competition wouldn't look like this. That said, we'd caution against reading platform count alone as proof of growth, since a high count paired with high churn, which we'll get to, is also consistent with a market still figuring out its winners.
129
AI girlfriend platforms actively tracked and tested
22%
now offer AI video generation, a newer feature category
18%
went dark, sold, or rebranded in a single re-audit year
Signal 2: how fast new feature categories are spreading
The clearest growth signal in our own data is feature adoption speed. AI video generation is the newest capability we track, and it's already reached 22% of the 129 platforms in our database, up meaningfully from when we first started tracking it as a rare, experimental feature. That kind of adoption curve, a feature going from a curiosity to a genuine minority-but-growing category within roughly a year, is a much more concrete growth signal than an unverified dollar estimate, because we watched it happen ourselves, platform by platform.
Signal 3: pricing experimentation and convergence
During a single re-audit pass, we found at least 39 platforms had added a genuine free tier that didn't exist during our original testing, and at least 28 had restructured a single flat price into multiple tiers. That level of active pricing experimentation across a big share of a 129-platform database is a strong indirect signal of real competitive pressure. Companies don't restructure pricing this aggressively in a market that's stagnant or shrinking. They do it when they're actively competing for the same, apparently growing, pool of potential subscribers.
At the same time, the market has converged hard around a specific price range. Averaged across the 85 platforms with a parseable price, the industry sits at $11.85 a month, and only 2 of 129 platforms charge premium prices. That convergence suggests a market that has found, through trial and error across many competitors, roughly what price point this category can actually sustain.
Signal 4: churn as a sign of consolidation, not decline
At least 23 of the 129 platforms in our database, about 18%, went dark, were sold and rebranded, or started redirecting to an unrelated product within roughly a year. We'd read this as consolidation rather than contraction. A market where 18% of participants exit within a year while the overall platform count stays around 129 or higher is a market where new entrants are actively replacing failed ones, which is a pattern typical of a still-young, still-forming category rather than a shrinking one. I go deeper into the specific patterns behind this churn in a separate piece on platforms that went dark, got sold, or rebranded.
Signal 5: where quality investment is concentrating
One more directional signal worth noting: quality investment isn't spreading evenly across the industry. Chat quality and pricing, the two categories that already scored highest in our testing at 3.26 and 3.30 out of 5, are also the categories that saw the least change during our re-audit. Voice, images, and support, the three weaker categories, saw much more visible movement, mostly in the form of a small number of platforms pulling meaningfully ahead of the pack rather than the whole industry improving evenly. That concentration pattern, strong platforms extending their lead in the hardest categories rather than the average simply rising, is itself a growth signal, just a different kind than a raw platform count.
What this suggests, and what it doesn't
Put together, our own data suggests a market that's still actively expanding its feature set (video generation adoption climbing), still experimenting aggressively with pricing (dozens of platforms restructuring in a single year), and still consolidating hard around its stronger competitors (18% annual churn). That's a reasonable, evidence-based case that this category is growing in relevance and sophistication, even without a dollar figure attached to it.
What it doesn't give us is a defensible revenue or user-count estimate, and we'd rather tell you honestly that we don't have one than hand you a number borrowed from a source we can't verify. If a platform or an article gives you a precise-sounding market-size figure with no visible methodology behind it, that's worth treating with real skepticism.
How we actually measure this over time
Everything in this article comes from our own testing database, re-audited periodically against the same 129 platforms using a consistent methodology. We track platform count, feature availability, pricing structure, and status (active, changed, or gone) directly, by revisiting each platform ourselves rather than relying on self-reported company data. You can read the full testing methodology, and see the complete current numbers on our data hub.
If what you actually want isn't an abstract growth narrative but a concrete answer to "which platform is worth trying right now," that's a different, more answerable question than market size, and it's the one our best AI girlfriend rankings are built to answer directly, using the same 129-platform testing data referenced throughout this article.
Further reading
Frequently Asked Questions
How big is the AI companion market?▾
We don't cite a specific dollar-denominated market size, since we can't verify third-party estimates' methodology. What we can confirm directly is that we track 129 active, competing AI girlfriend platforms.
Is the AI girlfriend industry actually growing?▾
Our own data suggests yes, directionally: AI video generation adoption reached 22% of platforms from a much smaller base, and at least 39 platforms added free tiers in a single re-audit pass.
Why don't you cite market-research firm estimates for this industry?▾
Because most platforms in this category are private and don't publish revenue or user data, making outside market-size estimates difficult to verify.
Does high platform churn mean the market is shrinking?▾
Not necessarily. At least 18% of platforms in our database went dark, were sold, or rebranded within a year, but overall platform count has stayed around 129 or higher, suggesting consolidation rather than contraction.
