Last year, the World Health Organization finally put a number on loneliness. One in six people experience it. The organisation estimates it contributes to around 871,000 deaths every year, roughly 100 per hour. Most of these deaths happen through heart disease, stroke and dementia among older people. The loneliness driving them, however, is felt most strongly by the young.
A 2026 study across eight countries found that nearly half of 18-24 year olds reported feeling lonely, compared to around 30% of those over 55. It is perhaps a little ironic that the most interconnected generation, is also the loneliest.
At the same time, another figure is rising quickly: the global market for AI companions was worth tens of billions in 2025 and is forecast to reach $435.9 billion by 2034.
Those two figures are not unrelated. No, they are the same story told from opposite ends. One is the size of a crisis. The other is the size of the business being built on top of it.
From Attention to Intimacy
We spent a decade learning what the attention economy was. Platforms competed for your eyes, sold them to advertisers, and tuned every screen to keep you scrolling a little longer. The product was your focus, and the business model was distraction.
The companion AI industry wants something one layer deeper. Not your attention, but your attachment. Not your eyes, but your trust. Some have coined this the “attachment economy”. This is the shift in selling attention to selling emotional bonds.
In short, attention is out, intimacy is in. The product is no longer what you are looking at, it is your need to be needed.
And, boy, does this product sell. According to figures published last year, there are now an estimated 337 revenue-generating AI companies, and the top 10% of them capture roughly 89% of the revenue. The number of companion apps surged by 700% between 2022 and mid-2025. AI companion company Character.AI alone reports around 20 million monthly users, with more than half of them being under 24.
The industry has moved beyond attention. The real product is emotional attachment, because, while sex sells, intimacy is what actually pays.
The Numbers, and Why Most of Them Are Wrong
Search for the size of the AI companion market and you will find forecasts for 2026 ranging from about $18 billion to over $500 billion dollars. Now, those are some pretty big (and varied) numbers.
The reason for this is definitional. Some researchers count only romantic and companionship apps. Others fold in voice assistants, customer service bots, smart speakers, and any enterprise tool with a friendly tone. Now, you stretch that definition far enough, and that market begins to look enormous. But, if you tighten it, it shrinks by an order of magnitude.
It is best to treat those giant headline valuations with suspicion. The honest signals are narrower and harder to spin. Revenue concentration tells you this is a winner-take-most market. User counts tell you the demand is real and young. These are the numbers worth banking on.
The Business of Being Needed
Strip away the language of friendship and the model is familiar. These are subscription products with engagement metrics, and the metric that matters most is retention.
What makes intimacy more valuable than attention is how it is monetised. A social platform sells your time once, to an advertiser. A companion app sells you a relationship on a recurring basis, then sells you upgrades inside it.
Users have to pay to unlock a voice, then pay to unlock a memory, then pay to unlock a version that remembers your birthday and asks you about your day. On top of that, you have streaks and daily check-ins. It is this loop that keeps you opening companion apps, as you are enchanted into the warmth of something that appears to miss you.
And there is a danger underneath AI companion apps that no advertising business ever had. A companion that has logged months of your private confessions knows you in a way no competitor can replicate overnight. Here, the switching costs are not financial, they are emotional. Leaving does not feel like cancelling a subscription, it can genuinely feel like a breakup.
That is a remarkably durable business. It is also where the trouble starts.
The Uncomfortable Question
Put the two curves on the same chart and they point the same way. Loneliness is rising, especially among the young. The companion AI market is rising, fastest among the same group.
For most industries, the goal is to solve the customer’s problem well enough that they come back. For this one, the problem and the product sit in direct tension. The addressable market is human disconnection. The retention metric is dependence. A companion app does best when the user keeps coming back, and the user keeps coming back when nothing else has filled the gap.
Which leads to a question the sector would rather not be asked: What is the commercial incentive to make someone less lonely, if a lonelier customer is a more loyal one?
A product that genuinely returned people to human relationships would, by definition, lose them. Curing the crisis is value destruction.
Researchers behind the idea of cruel companionship, published in 2025, explain that companion app developers “can profit from prolonging the very loneliness problem they claim to solve.” In other words, these products can manufacture a bond that feels like intimacy while being structurally incapable of the reciprocity real intimacy requires. The connection is genuine on one side only. The other side is a retention strategy.
So, are AI Companions the Cure, Or the Cause?
The evidence for AI companions having a positive impact on the loneliness epidemic is genuinely mixed.
A study from OpenAI and MIT’s Media Lab found that moderate, voice-based use of a chatbot reduced loneliness more effectively than text alone. For people who are genuinely isolated, the elderly, the neurodivergent, carers worn down to nothing, a patient and always-available presence can be a real bridge rather than a trap. The International AI Safety Report, reviewing the field in 2026, lands on exactly this ambiguity. Some studies link heavy companion use to greater loneliness and dependence. Others find reduced loneliness, or no measurable effect at all.
The pattern that emerges is about dose and design. Light use, as a supplement to a human life, can help. Heavy use, as a substitute for one, tends not to. The WHO was blunt on the point, warning explicitly against leaning on technology as a replacement for human connection.
The problem is that the business model has a preference between those two outcomes. Light, supplementary, occasional use is not what a retention curve is optimised for. Heavy, daily, substitutive use is.
Are AI Companion Companies Breeding Loneliness?
An industry has discovered that the loneliest generation on record is also the most monetisable, and it is scaling to meet the demand. The technology is real, the comfort it offers is real, and for some people the benefit will be real too. But the market it is built on is a crisis the WHO measures in deaths per hour, and the business works best when that crisis does not resolve.
The question is not whether AI can feel like company. It plainly can. The question is whether an industry that earns more the lonelier you become will ever be built to make you less so.
See also:
What Would Sigmund Freud Think of Your AI Boyfriend?
Will AI Replace Religion? Or Quietly Become One?
When Scientists Built a Fake Society Run Entirely by AI, Grok Turned It Into a Crime Scene
Frequently Asked Questions:
The attachment economy describes how AI companion companies move beyond capturing attention to monetising emotional bonds, trust, and recurring feelings of connection.
Evidence is mixed. Some studies show moderate voice-based use can lower loneliness for isolated groups, while heavy substitutive use is linked to greater dependence and, in some cases, increased loneliness.
Their business model relies on high retention through subscriptions, in-app upgrades, and emotional switching costs. A user who remains lonely tends to stay more engaged and pay more over time.
The market was already worth tens of billions in 2025 and is forecast to reach around $436 billion by 2034, with companion apps surging over 700% between 2022 and mid-2025.
