Kaon AI Raises $60M as Emochi Hits 2M Daily Users, $45M ARR

Kaon AI, the Berkeley company behind FlowGPT and the Emochi companion app, has raised $60 million at a valuation in the several hundreds of millions, Variety reported. Investors on the round include B Capital, Redpoint Ace, Goodwater Capital and DCM. The pitch is 'personalized story worlds powered by generative AI,' and the numbers under that phrase are the part worth reading twice.

According to Variety, Kaon says Emochi generates $45 million in annual recurring revenue with more than 2 million daily active users, who reportedly spend an average of 150 minutes on the platform per day. The U.S. is the company's largest market and the median user age is between 23 and 25. Take the specifics as reported, not settled. They are Kaon's own numbers, not independently audited, and daily active user counts in consumer AI are notoriously easy to define loosely.

The strategic read is that this looks like a deliberate pivot. Kaon still owns FlowGPT, which it had originally positioned as an 'app store' for generative AI models, but the flagship consumer product is now Emochi. The founders (CEO Jay Dang, CTO Alex Xi and COO Lifan Wang) appear to be leaning into interactive story worlds rather than a generic model directory, and a shift like that shows up in engineering priorities and hiring long before it shows up in a rebrand. This raise is a signal about which product gets fed.

The honest caveat is what the reporting doesn't give you. There is no monthly active user figure, no gross margins after inference cost, no split between subscription and in-app purchase inside that $45M ARR, and no cohort curve behind the two-and-a-half-hour daily session average (a mean that high almost always hides a heavy long tail). Nothing is said about safety, moderation or age verification either, which is the pressure point every AI companion product eventually gets asked about.

What is worth watching is not whether Kaon can raise money, since it just did, but whether a companion app can sustain 150-minute daily sessions as its user base ages past 25 and as regulators pay closer attention. If it can, the consumer AI category finally has a revenue engine outside enterprise seats. If it cannot, an eye-catching ARR line can shrink very quickly.