AI Social Apps Have Demand, But Monetization Is Hard

Putting my take out there first. Big tech pausing AI social doesn't mean the demand is dead; it just means the business model for generic companion chat is getting blocked by regulators and platform rules.

The evidence is straightforward too. Japan's loneliness economy is mature, with dating sims and anime fully developed, yet overseas AI social hasn't truly captured this market.

a16z says AI is too cheap—I think they're half right. Generating characters is indeed cheap now, but getting users to keep paying requires boundaries, gameplay mechanics, payment systems, and compliance. In Grok's line, Ani uses companionship, flirting, and affinity mechanics to push Japan into being the second-largest revenue market, proving willingness to pay isn't the issue.

The problem lies in valuation logic. If early-stage projects rely on edgy content to make money, no matter how good retention looks, it's still a platform risk. I know a founder working in a similar direction; team execution is key, but what they fear most is channels, regulators, and model providers tightening up all at once. Model anthropomorphism actually ranks lower in their concerns. The delisting of Weiban and the rectification of Xingye are cautionary tales.

The judgment is simple. The core barrier for AI social is risk boundaries; model anthropomorphism comes second. What's investable are products with vertical scenarios, clear monetization, auditability, and those that avoid crossing general assistant red lines. Pure chatbots? Don't touch them. That's it.