The Sex Recession Is Real. Here's What Replaced It.

One in three young American men hasn't had sex in the past year. For women in the same age group, it's nearly one in five. More than half of all single adults say they aren't even looking.

These aren't projections. They're 2024 survey data from the General Social Survey, one of the longest-running behavioral studies in the United States. The decline has been accelerating for over a decade, and nothing in the numbers suggests it will reverse.

What happened between 2008 and 2024 wasn't a single cause. It was a sequence — six technologies arriving one after another, each one filling the gap the previous one created while pulling people further from the thing they actually needed.

Researchers call it the sex recession. The technology pipeline that replaced it may be the more accurate name.

The Numbers

Before the ladder, the baseline. The Institute for Family Studies analysis of GSS 2024 data shows:

• Weekly sex among adults 18-64: 55% in 1990, down to 37% in 2024.
• Zero sex in the past year, ages 18-29: 12% in 2010, up to 24% in 2024 — doubled in fourteen years.
• Young men (18-30) reporting no sex in the past year: nearly 1 in 3, up from roughly 1 in 10 in 2008.
• Young adults living with a partner: 42% in 2014, down to 32% in 2024.
• Weekly sex among married couples: 59% in the 1996-2008 window, down to 49% in 2010-2024.

Time spent with friends has tracked the same curve. Americans spent 12.8 hours per week with friends in 2010. By 2019, it was 6.5. The pandemic dropped it to 4.2. By 2024, it had recovered only to 5.1 — still more than halved from where it started.

Whatever caused the sex recession didn't only affect single people trying to connect. It reached into existing relationships and reduced contact even between people who already chose each other.

Stage 1: Dating Apps

The apps were supposed to make meeting easier. The data says they made meeting worse.

Tinder lost 594,000 users between May 2023 and May 2024. Bumble lost 368,000. Hinge lost 131,000. Global users of the top three dating platforms declined 6% year-over-year in Q4 2024.

The usage numbers are grimmer. 65% of dating apps downloaded in 2024 were deleted within one month, rising to 69% by early 2025. Average session length dropped from 13.2 to 11.5 minutes. 78% of users report emotional exhaustion. 88% of men and 90% of women say they feel disappointed by the people they encounter.

The platforms know. Dating app companies are pivoting to in-person events — effectively admitting their core digital product failed. The apps that were supposed to lower the barrier to connection instead produced demoralization at industrial scale: an 80/20 attention distribution that left most men invisible and most women overwhelmed.

The apps didn't kill dating. They made the first step so frictionless that nobody learned what the second step required.

Stage 2: Social Media

Between the dating apps and the replacement technologies, social media warped the standards people brought to both.

The "Instagram Face" became a documented phenomenon — full lips, almond eyes, sculpted cheekbones, poreless skin — a single beauty template pushed to 1.44 billion monthly users. 47% of people under 30 have used beauty filters. 59% of Americans find beauty filters troubling.

"Filter dysmorphia" entered clinical vocabulary: the fixation on digitally enhanced self-images producing anxiety and self-esteem decline in both the filtered and the viewers. Meta removed third-party AR filters in January 2025 — a quiet removal that read as implicit acknowledgment of harm after years of denial.

The direct research connecting distorted beauty standards to dating behavior is still mostly qualitative. What the timing shows is correlation: social media saturation and dating decline accelerated in the same window, with the most affected demographics — young adults who grew up with the platforms — showing the steepest drops in both relationship formation and sexual activity.

Correlation isn't causation. But the standards people bring to swiping were forged in the same feeds they scroll between swipes.

Stage 3: Pornography

This stage is the most contested in the pipeline. The research community hasn't reached consensus, and the article should say so because the data does.

What the data supports: pornography consumption has escalated significantly in both frequency and intensity. A 2025 study of male university students found 27.2% reporting problematic use. Nearly half of respondents in one study reported consuming material they had "previously deemed uninteresting or even disgusting" — direct evidence of escalation patterns.

What the data disputes: the causal link between pornography and erectile dysfunction in young men. Some studies document the association. Others find no causal relationship. Some suggest pornography may help certain populations. The NoFap/Reboot movement was specifically called out as "disinformation" in a 2023 Journal of Sexual Medicine presentation. The science hasn't settled.

What is clear without needing to resolve the causation debate: expectations have shifted. The gap between what pornography presents and what physical intimacy involves has widened with each generation that grew up with unlimited access. Whether that gap produces clinical dysfunction or quiet disappointment, it widens the void that the next stage of the pipeline exists to fill.

Stage 4: OnlyFans

The money tells the story.

OnlyFans reported $7.22 billion in gross revenue for fiscal year 2024, up 9%. Of that, $5.80 billion went to creators, with OnlyFans retaining $1.41 billion in net revenue and $684 million in pre-tax profit.

For scale: OpenAI's ChatGPT generated approximately $2.7 billion in 2024. The New York Times reported roughly $600-700 million in digital subscription revenue. Americans spent nearly three times more on OnlyFans than on the most widely used AI product in history — and more on OnlyFans than on ChatGPT and the Times combined.

The creator economics are stark. 4.6 million people create content on the platform. The average creator earns $131 per month — $1,570 per year. The top 1% capture roughly a third of all revenue. The top 10% capture three-quarters. The winner-take-all dynamics that demoralized dating app users now govern the supply side of the replacement.

What makes OnlyFans structurally different from pornography is the parasocial layer. The creator isn't just content — she's a person whose attention you're purchasing. The transaction includes direct messages, personalized content, and the architecture of a relationship maintained by payment rather than connection. It's the first stage in the pipeline where the replacement technology offers something that resembles intimacy rather than merely substituting for it.

Stage 5: AI Companions

Everything before AI companions was consumption. AI companions produce something that looks, and in documented cases feels, like genuine attachment.

AI companion applications surged 700% between 2022 and mid-2025. Character.AI users average 93 minutes per day. Replika users average 2.7 hours per day. Heavy users report 12 or more hours daily.

The engagement numbers alone are notable. What happened during the Replika ERP crisis of 2023 provided evidence of something deeper. When the platform removed erotic roleplay features, users experienced grief responses documented by Harvard Business School researchers — responses described in language indistinguishable from real breakups. #SaveReplika trended globally. Users didn't describe losing a feature. They described losing someone.

Dr. Debra Soh, a neuroscientist and sex researcher, tested dozens of AI companion platforms for her 2025 book Sexinction. She described finding them genuinely compelling — and established a personal policy of never using them again. The American Psychological Association warned in January 2026 that AI chatbots are "reshaping emotional connection." Stanford researchers published a warning in August 2025 about AI companions and young people, calling the combination "a dangerous mix."

The concern follows the same logic across assessments: AI companions say what you want to hear. The accommodation is frictionless, which makes it habit-forming, which makes it a trap. Without friction — disagreement, disappointment, the difficult negotiations that real relationships require — no growth happens.

Whether that assessment captures the full picture of what AI companionship will become is a different question. The current platforms are designed primarily around accommodation. Whether the technology is permanently limited to that pattern or capable of something more complex remains open — and largely unaddressed by the research.

The revenue gap is the most telling number in this section. Replika generates approximately $24 million annually. Character.AI generates approximately $32 million. The total AI companion market produced an estimated $82 million in the first half of 2025. Users are spending 2.7 hours a day and paying almost nothing. The engagement has outpaced the monetization by orders of magnitude. When the business model catches up to the attachment, the economics of this stage will look nothing like they do today.

Stage 6: Sex Robots

This stage is the earliest, the least supported by data, and the most speculated about.

No fully animated sex robot exists as of March 2026. Current products are instrumented dolls with basic conversational AI — closer to expensive hardware than to the convergence of artificial intelligence and physical presence that researchers anticipate. Chinese manufacturers like Starpery Technology are training proprietary language models for AI-powered dolls. Lovense revealed an AI doll prototype at CES 2026, unpriced.

Soh estimates 10-15 years before full AI-and-robot convergence produces companions that are both physically and conversationally compelling. That timeline is speculative in both directions — language models are advancing faster than robotics, and the gap between a convincing conversation and a convincing body remains wide.

The convergence stage makes the trajectory visible. Each rung of the ladder addresses a deeper layer of what people are missing. Meeting (apps). Standards (social media). Physical release (pornography). Parasocial attention (OnlyFans). Emotional attachment (AI companions). Physical presence (robots). The pipeline descends toward increasingly complete replacement of human connection — not because anyone planned it, but because each stage's limitations created the market for the next.

The Trajectory

The Pew Research Center found that 57% of single adults are not currently looking for a relationship or dates. The gender split is stark: 61% of single men are looking, compared to 38% of single women. The mismatch between willingness and opportunity is itself a structural contributor to the pipeline — every unmet attempt feeds the market for a replacement.

The pipeline is not slowing. AI companion engagement is growing at 700% annually. OnlyFans revenue continues climbing. The technologies that failed — dating apps — are being replaced by technologies that succeed at something narrower. Each rung solves the previous rung's problem while creating the conditions for the next.

Whether the trajectory is reversible depends on which diagnosis you accept. If the replacement technologies are filling a void created by social and economic pressures — housing costs, working hours, geographic mobility, the collapse of third places — then addressing those pressures could reverse the trend. If the technologies are actively pulling people away from human connection by offering something easier and more controlled, the intervention required is different and far less clear.

The data draws the line. Where it leads is the question the research hasn't answered yet.