The AI Turnaround Gap
New survey data puts current U.S. dating-app use at 6% of adults. Days later, Match Group and Bumble report earnings that show why their AI bet hasn't closed that gap yet.
Match Group reports second-quarter earnings on August 4th, and Bumble follows a day later, on the 5th.[5][3] Analysts are not expecting surprises: modest revenue, fewer paying subscribers, and executives pointing to new AI features as the fix. In the same window, new survey research from SSRS found that only 6% of American adults are currently using an online dating site or app at all.[1] The apps are not just losing payers one quarter at a time. They are losing the argument that most adults belong on them in the first place.
Who's Actually Still There
The SSRS numbers are worth sitting with. 37% of U.S. adults say they have used an online dating site or app at some point in their lives, but current use drops all the way to 6%.[1] That gap between "ever tried it" and "still doing it" is the real story, and it isn't evenly spread. About half of adults aged 18-29 and 30-49 have used online dating, versus 20% of adults 50 and older, and the same pattern holds for current use: 10% and 8% respectively, against just 2% for adults over 50.[1] LGBTQIA+ adults use these platforms at nearly double the rate of everyone else — 63% have ever used one, including 14% who are current users, compared with 34% and 4% among non-LGBTQIA+ adults.[1]
Even within that shrunken base, one brand still dominates. Among adults who have ever used an online dating site or app, 48% say they've used Tinder — more than any other platform, with Plenty of Fish and Facebook Dating tied for a distant second at 30% each.[1] Put the two findings together and the picture sharpens: dating apps are no longer a mainstream utility that most adults dip into. They are a concentrated habit, held by younger adults and LGBTQIA+ adults, run through one brand more than any other.
The Earnings Match the Survey
Bumble's first-quarter 2026 numbers show the same contraction in dollar form. Revenue fell 14% year-over-year to $212.4 million, and paying users were down 23% year-over-year, with the sequential decline described as worsening.[2] Bumble's own guidance for the second quarter — $205 million to $213 million — implies the slide continues rather than reverses.[2] Analyst estimates ahead of the August 5th report put Q2 revenue near $210.46 million and EPS at $0.25, in line with that guidance; the prior quarter's actual EPS of $0.34 beat consensus, but net margin was a deeply negative -72.04%, even as return on equity came in positive at 32.66%.[3]
Match Group's most recent quarter told a similar story from the other big incumbent: reported EPS of $0.612 came in under the $0.68 the market expected, and the consensus forecast for the quarter reporting August 4th is $0.65 — a cautious number, not a rebound.[4] Both companies confirmed their August reporting dates weeks in advance, standard corporate housekeeping that nonetheless puts two of the industry's largest names on record, in the same week, explaining why fewer people are paying them.[5][3]
The AI Pitch Is Ahead of the AI Result
None of this is happening without AI on the table. Bumble has spent the past year on what one analysis called an "aggressive AI" push under a founder-led turnaround — and still posted the 14% revenue decline and eroding market share described above, with competition from Hinge cited by name as the pressure point.[2] That is the honest version of where AI-in-dating stands at the incumbent level: real investment, real feature launches, and a user base that keeps shrinking anyway. The AI in question mostly does the same job the swipe interface always did — hold attention inside a feed — just with a smarter icebreaker or a safety filter bolted on. It is not changing who is doing the deciding, or how much of a stranger's actual character a user sees before they decide.
The apps are betting on AI to reverse a decline that, so far, the AI hasn't reversed.
A Narrower Bet on a Narrower Task
AISURU starts from the SSRS numbers rather than around them. If the people still willing to date through an app are a smaller, more self-selecting group — younger, disproportionately LGBTQIA+, already skeptical of a feed that hasn't worked for them — then the product worth building isn't the one that maximizes swipes per session. It's the one that treats a user's time as finite from the start. AISURU asks for five essays, 300 words minimum each, before it does anything else. An AI reads them and extracts more than sixteen personality traits, then scores compatibility across four weighted dimensions: lifestyle at 35%, emotional depth at 30%, complementary differences at 20%, and shared values at 15%. Only pairings that clear a 65-point threshold are ever shown, and they arrive once a day, not as a feed to refresh.
It's worth being precise about what that AI does and doesn't do, because the distinction is the whole design. AISURU's model is analytic: it reads what a person actually wrote and scores fit against another person's answers. It does not generate a profile, write a bio, produce a photo, or send a message on anyone's behalf. It never pretends to be the user. That is a narrower, less flashy use of AI than an incumbent's chat assistant or auto-generated opener — and it will not, by itself, reverse the usage numbers SSRS just published. But the earnings this week make the counterfactual worth asking: after a year of AI investment at the biggest apps in the category, the paying user base still shrank. A model built to do less, more carefully, is at least a different bet than doing the same thing with a smarter coat of paint.