Resolves to YES if the top male companion is sustainably more popular (as in, is verified to serve more queries, more tokens or more DAUs or MAUs, in that order, based on which stats are known, or to any common sense clear indication that this isn't close) than the top female companion by EOY 2025.
Resolves to NO if the female companion is more popular the same way.
If we do not have an official known answer by EOY 2025, resolves to my percentage estimate that this is true, probably the market price.
If Grok does not release a male companion by 11/1/25 this resolves N/A (or if that is disallowed, to fair market price).
[I will not be answering questions about possible corner cases unless this market gets huge, otherwise I'll do whatever seems fair if they happen.]
1,000
3.00My model is roughly
- female and male userbase grow correlated but not perfectly so (i.e. social graph is gender-correlated) ; there is a natural pressure towards 1:1 , but it's not huge
- in particular, character.ai started at and is at 1:1 roughly, while Replika seems to be at 2:1 to 4:1 male:female userbase despite also (apparent?) neutral marketing
- so far Grok has been targeted heavily at men, male userbase will grow quite a lot now shifting male:female more towards I guess oom 15:1 (female users using female companions just isn't that frequent, and again, social graphs argument for how grok adoption will work)
- when Grok shifts, it will get a push towards 1:1 again, but since Replika seems stable away from 1:1 I don't expect this to get us far ; in particular, Grok has done a gendered marketing campaign unlike Replika (afaik) so it's gonna be worse ; otoh Grok relies only partially on word-of-mouth recommendations, while awareness of the offer is done by pushing the feature into everyone's face directly
As a result I've changed my mind from expecting some sort of equlibrium similar to romance novels to be reached within 2025, towards sth like... maybe the equilibrium doesn't apply for AI companions (not sure why) and definitely that convergence is slower than a few months, even during exponential growth phases.