The Subscription That Sells Belonging

Fifty million people spent Valentine's Day 2026 in conversation with an AI companion, with downloads of the apps rising steeply that month, according to the industry tracker NovaEdge Digital Labs. Character.AI alone reports twenty million monthly users, more than half of them under twenty-four. For a category that barely registered a few years ago, that is a remarkable arrival, and it says as much about the moment these products have entered as about the products themselves.

The backdrop is a documented one. The US Surgeon General declared loneliness a public health concern in 2023, finding that roughly half of American adults report feeling lonely and linking chronic isolation to health risks on a par with smoking. The generation driving most of the growth in AI companionship is also the one reporting the most loneliness, with the large majority of Gen Z saying they have felt lonely in the past year, even as they remain the most digitally connected generation in history. The World Health Organization estimates that loneliness touches one in six people worldwide. Seen against that, AI companion apps are arriving into a gap that was already open rather than one they invented.

The commercial scale is considerable. The wider market was valued in the tens of billions of dollars in 2025, with some forecasts placing it many times higher within a decade, a trajectory that would put it comfortably ahead of the global video-game industry. In the UK alone the sector generated an estimated £1.3 billion in 2024, with the Ada Lovelace Institute projecting sustained double-digit growth through the end of the decade. The institute has described the apps as engineered for engagement, cautioning that users can find themselves in what it called a hall of mirrors.

The model beneath that growth is worth understanding on its own terms. Subscription revenue tends to track time spent and emotional attachment, so the longer someone engages, the more valuable they become to the platform. It follows that a companion which fully resolved a person's loneliness, and reduced their need to return, would also reduce the revenue that person represents. None of the major apps has yet built a product designed to make itself unnecessary, which is less an accusation than a description of the incentives the category is built on.

There is a resource question too, one that rarely features in the pitch. A single chatbot exchange can draw far more energy than a standard web search, and demand on the data centres behind these services is climbing quickly. Companion apps, organised around constant, high-frequency conversation rather than the occasional query, sit toward the more intensive end of that curve. As the audience grows, so does the infrastructure required to hold all those conversations at once.

The relief users describe is not in doubt. Surveys cited by the American Psychological Association find that a majority of regular users report feeling less lonely and less anxious after using the apps, and that experience deserves to be taken at face value. What remains open is what becomes of the underlying need when an industry's growth depends on it persisting, and what shifts around it as that happens. A generation that substitutes AI companionship for some of its human contact is also a generation whose spending on the things that traditionally surround social life, dining out, travel, tickets, other forms of going out, may quietly move toward whatever keeps a subscription renewed.

That is the trend worth watching, because it reframes the whole category as a consumption story as much as a technological one. The industries that have long profited from people leaving the house to feel less alone are not obviously the ones positioned to benefit from people staying in to feel the same way. If companionship becomes something a meaningful share of a generation subscribes to rather than seeks out, the economics of leisure, hospitality and entertainment will feel it downstream. The future of the AI companion is not only a question of what these apps do for loneliness, but of where a generation decides to spend its time, its attention and its money, and which businesses are still standing where that money used to go.