Replika founder’s Wabi hits $1.25B value, betting AI works better in groups

Wabi, the personal AI assistant startup founded by Replika creator Eugenia Kuyda, has raised $55 million at a $1.25 billion valuation. IVP led the round, according to the company’s release. Bloomberg reported the deal the same day.

A round led by IVP with Alexa Fund money

Besides IVP, the investors named in the PR Newswire release are Andreessen Horowitz, Forerunner and the Amazon Alexa Fund. Several angels and founders also took part: Tobi Lütke, Marc Benioff, Marissa Mayer, Dick Costolo and Alex Pall.

Wabi says the round followed strong interest after its recent public launch. The money is meant to fund further product development and expansion. The company is based in San Francisco, and Kuyda is its founder and CEO. She previously founded Replika, the AI companion platform.

One agent, mini-apps and shared spaces

Wabi describes its product as an AI agent combined with personalized mini-apps and shared spaces. The idea is that people use AI together, rather than one person talking to a chatbot alone. Bloomberg’s summary puts it as software that lets groups interact with AI agents to coordinate activities and make collective decisions.

The company lists some uses it has seen since launch: friends finding music, group fitness challenges, couples planning dates and individuals building routines. Sample mini-apps include shared restaurant lists, meal plans that turn into grocery lists, trivia games and book club trackers. Wabi also mentions help with family schedules, homework, booking appointments and trip planning.

The app is available now at wabi.ai, on iOS, Android and Mac. Windows is listed as coming soon. The release does not mention pricing, a free tier, user numbers or supported regions.

Kuyda says in the release that Wabi aims to become “an operating system for life.” IVP general partner Alex Lim said Kuyda “has a unique sense for where AI is headed.”

What the announcement leaves unanswered

The release gives a headline valuation but little to test it against. It says nothing about how many people use Wabi, how many are active, whether the app charges anything, or what revenue it earns. “Strong early demand” is the company’s own description and has not been independently verified. The same goes for the usage examples, which come from Wabi itself.

The size of the round also needs context. At $55 million, it is small next to other AI deals reported this week. Our earlier coverage noted that FuriosaAI raised $620M at a $2.9B valuation, and that Cognition is fielding offers at $70B-$80B, only weeks after a $48B round. Against those, Wabi’s $1.25 billion value rests on a thin disclosed base: a $55 million raise and a recent launch. Investors appear to be paying for the founder’s track record and the shared-AI concept as much as for numbers the company has published.

The Alexa Fund’s presence is notable for an assistant product, but the release does not say whether the investment comes with any commercial arrangement with Amazon. It also does not say how Wabi’s agent works under the hood, which AI models it uses, or how it handles the personal data that family, couple and friend groups would share inside it. Those are the questions group products raise most, since several people’s information sits in one space.

The concept is also not unique. AI agents increasingly connect to outside tools and data through shared standards, as our explainer on the Model Context Protocol describes. Wabi’s difference, on its own account, is that the social setting is the starting point rather than an add-on.

The test is whether groups keep using it

Wabi has money and a name investor list, and it is already in app stores and on the Mac. What it has not shown is whether the shared-space model keeps people coming back once the novelty fades. The company has not published retention figures or user counts, and the release outlines no timeline beyond the coming Windows version. Readers who want to judge it can try the app directly at wabi.ai.