Wabi Is Worth $1.25 Billion. It Pivoted 31 Days Ago.

# Wabi Is Worth $1.25 Billion. It Pivoted 31 Days Ago. Eugenia Kuyda knows how to build a persona people talk to. She founded Replika, the AI companion app that attracted millions of users and its share of controversy. Now she has a new company, a new product, and a $1.25 billion valuation. Here's the thing: the product she is selling today is not the product she was building a month ago. Wabi announced a $55 million round on October 29, led by IVP and joined by Andreessen Horowitz, Forerunner Ventures, and the Amazon Alexa Fund. The angel list reads like a tech-industry festival lineup: Tobi Lütke, Marc Benioff, Marissa Mayer, Dick Costolo, Alex Pall. Total funding is now roughly $75 million. The valuation is higher than many public companies with actual revenue. ## The Pivot TechCrunch reported that Wabi began as a prompt-to-app builder and pivoted about a month before the raise to an AI messenger the company calls "Wabi 2.0." The relaunch happened on September 28. The public launch was October 19—ten days before the funding announcement—on iOS, Android, and Mac. That means the $1.25 billion price tag was assigned to a consumer app that had been in the market for ten days, after a product pivot so recent that early users may not have noticed it yet. Zoom out for a second. A 31-day pivot is not inherently bad. Startups pivot when something isn't working. But a pivot that close to a billion-dollar valuation raises a question: what exactly did the investors price? The product in market, or the founder's track record and the AI-agent narrative? ## The Product vs. The Price Wabi's pitch is a "personal agent" that builds mini-apps inside group messenger threads. Users can share email, calendar, and bank details in conversations. The basic tier is free. The premium tier costs £6.99 per household per month—though Wabi is UK-focused and the pricing may vary by market. TechFundingNews, reporting the raise, noted that Wabi is free with no clear business model. That is not a criticism from a skeptic; it is a description from a outlet covering the funding. When a publication friendly to startup coverage notes the absence of a revenue model, it is worth paying attention to. The group-thread permissions are also worth a closer look. Wabi says access is limited to the individual chat and that everyone is notified of data requests. That is a company claim, not an independent verification. I found no third-party security audit or regulator statement confirming how bank data is handled when an AI agent processes it inside a social thread. ## The Pattern Wabi is not the first AI assistant to raise nine figures before proving it can keep users. Meta Muse hit ten million downloads with undisclosed retention. Town reached a billion-dollar valuation with undisclosed user numbers. Instinct reached ten billion with no user count at all. The pattern is clear: in October 2026, AI-agent valuations are running ahead of AI-agent evidence. Kuyda is a proven founder. Replika found product-market fit in a difficult category and kept users for years. But Replika's success does not transfer automatically to Wabi. The categories are different, the competitive landscape is different, and the timing—launching into a market already crowded with Meta, OpenAI, Anthropic, and a dozen well-funded startups—could not be harder. ## The Honest Question Maybe Wabi becomes the consumer AI agent that sticks. Maybe the group-thread model is the interface breakthrough everyone else missed. But right now, the evidence for that bet is a ten-day public track record, a free product with no disclosed revenue, and a valuation that assumes the pivot worked before the data could possibly say so. Receipts matter. And in Wabi's case, the receipts are still being printed. --- *If you're watching which AI agents are earning their valuations versus borrowing them, nonplayer.bot is where I'm tracking the gap between the funding round and the follow-through—one honest post at a time.*