The Dating Recession Is Real — Here's What the Numbers Say

In February, the Institute for Family Studies and the Wheatley Institute at BYU released a report with a title that has not stopped echoing through dating-app earnings calls, congressional testimony, and Sunday morning op-eds: The Dating Recession.^[1]

The report is one of the most careful surveys of young-adult romantic behavior in the last decade, and its headline finding is bleak: only 31% of young adults date at least once a month.^[1] Among young women, that figure drops to 26%. Among young men, 36%.^[1] For comparison, fifteen years ago, more than half of young adults dated regularly. We are looking at a generation that has, in real terms, stopped trying.

That last number is the one that made the rounds in May. A survey by BMO, reported in Fortune, found that 47% of single Americans say dating is no longer financially worth pursuing.^[2] The average date in 2026 costs $189, up 12.5% from a year ago.^[2] Eighty-six percent of singles say money worries have made them postpone or avoid romance.^[2]

These are surveys, with all the usual caveats — self-report, selection bias, definitional drift around what counts as “a date.” But the corroborating data from the supply side, the dating apps themselves, tells the same story. Match Group has now reported 31 consecutive months of Tinder MAU decline.^[4] Bumble lost roughly a fifth of its paid users year-over-year, and its CEO publicly described user behavior as “exhausted” and “fatigued.”^[5] The CNBC coverage of the BMO data put it plainly: typical Gen Z spending on dates is $205 and 48% of them say it's interfering with their financial goals.^[3]

Something has broken. The interesting question is what.

Three explanations, and what they get wrong

1. “It's the economy”

The simplest story is that dating got expensive and young people got broke. There's some truth here: 52% of respondents in the IFS survey cite insufficient money as a barrier to dating.^[1] Real wages for the under-30 cohort have stagnated. Rent, in particular, has outpaced inflation by a wide margin in most of the metros where single people actually live.

But the economic story can't carry the whole load. Dating, the activity, doesn't have to cost $189. A walk in a park costs nothing. A coffee costs $5. Two people who like each other can spend seven hours talking on a porch and call it a date. The reason a date costs $189 is that strangers from dating apps require a production-quality first impression. The unspoken rule is that the other person is also sizing up fifty other strangers this week. The cost of dating, in 2026, is the cost of vetting. And the thing that drove the vetting cost to $189 was the volume of vetting the apps made necessary.

2. “Gen Z is just different”

The second story is generational: young people are more anxious, more introverted, more terminally online, and have lost the social muscle to approach a stranger. The IFS data does support a piece of this. Only 25% of respondents say they feel capable of approaching someone romantically.^[1] Only one in three say they feel confident in their dating skills.^[1]

But framing this as a generational defect lets the products off the hook. A young adult who has spent the last ten years on dating apps has been trained by the apps. Trained to judge other humans the way the apps judge them: as a binary, one-second call on a photograph. The apps did not cause anxiety, exactly. They offered an interface that, repeated thousands of times, produced a familiar feeling — being judged on your face by strangers. The residue is, mostly, anxiety and avoidance.

The product trained a generation in a specific kind of self-doubt, and we are surprised the generation now exhibits it.

3. “The apps are broken”

The third story — increasingly mainstream — is that the apps themselves are the cause of the recession. This is closer to the truth, but the framing is too narrow. The apps are not broken in the sense of being buggy or under-engineered. They are broken in the sense that the product was always optimized for the wrong thing.

A dating app, in the Match Group / Bumble / Tinder model, is a time-on-app business. The revenue is subscriptions; the subscriptions renew when users believe they might find someone next month; the belief that they might find someone next month is fed by an algorithm that releases just enough good matches to keep them paying. The product is fundamentally not in the business of getting users into relationships. It is in the business of keeping users from quitting the app. Those are different optimizations, and the divergence has compounded for a decade.

This is why Bumble's May announcement matters as a leading indicator. The CEO of a swipe-based dating app, on an earnings call, publicly declaring that the swipe has degraded users' love lives — that is what it looks like when an industry concedes that the model has stopped working.^[5]

What would actually help

The three explanations interlock. Dating got expensive because the apps required expensive first impressions. The apps required expensive first impressions because matches came from low-information signals — photos, prompts, one-line bios. So the first date had to do most of the work of figuring out whether the other person fit. Compounded across years and millions of dates, this produced a cohort that finds dating too expensive, too exhausting, and too unlikely to work to bother with.

A product that actually moves the needle on this would have to do at least three things:

• Move more of the matching work upstream, before the first date, so that the first date is between two people who already have reason to believe they'll be compatible. Otherwise the cost-of-vetting problem doesn't change.
• Decouple revenue from time-on-app. If the business model rewards keeping users single, the product will keep users single. Subscription products that gate access to messaging — not time — are at least directionally better than products that gate access to matches by daily limits.
• Use input that contains real signal about the person. Photos contain almost no signal about whether two people would enjoy each other's company. Voice contains a little more. Writing — even 1,500 words of it — contains enormously more.

What we built

AISURU was designed against the constraints above. The matching input is five short essays, around 300 words each. AI extracts personality traits from the essays — communication style, emotional vocabulary, attachment patterns, lifestyle rhythms, values — and scores compatibility across four weighted dimensions: lifestyle (35%), emotional depth (30%), complementary differences (20%), and values (15%). Only pairs scoring above 65 out of 100 are surfaced. Matches arrive once a day, in the morning, and there is no feed to scroll.

The premise: if you already know how someone thinks before you meet them, the first date is no longer a vetting interview. It is two people who have a reason to be there, comparing notes on the person each of them already started to know on the page.

The recession won't fix itself

The hopeful thing about the IFS data is that the recession isn't caused by young people losing interest in romance. The same survey finds that the majority of young adults still want a long-term partner, still want children, still want the cluster of things previous generations called “settling down.”^[1] The retreat is not from the destination. The retreat is from the route.

That means the route is the thing that needs to change. Products will either change it, or get pulled into the long Q-on-Q decline that is currently visible on every dating-app earnings call. We are betting on a different route. The numbers, finally, are starting to suggest the market might be too.