A famous founder, an unfamous idea, and a very familiar bet
Here is a fun feature of venture capital: you can raise $8 million at a $35 million valuation for a product nobody has seen. Dazzle AI did exactly that in December 2025. The company will tell you it is building a consumer AI app - a "next-generation personal assistant" - and then it will politely decline to tell you much else. What it has, instead of a demo, is a founder.
That founder is Marissa Mayer, who was Google's twentieth employee and its first female engineer, and later the CEO of Yahoo. Both of those facts are the kind of thing that lets you raise money on a promise. The promise here is specific and, in a crowded field, almost contrarian: that the next frontier in artificial intelligence is not the foundation model but the application built on top of it - the part regular people actually touch.
This is a sensible-sounding thesis, and it is also a convenient one. Mayer is not going to out-spend OpenAI or Google on training runs, and she isn't trying to. Dazzle's whole argument is that the interesting problem has moved up the stack, to design and delight, which happen to be the things Mayer has spent a career on. When your competitive advantage is taste, you argue that taste is what matters.
Dazzle did not appear from nowhere. It is the successor to Sunshine - originally Lumi Labs - the contact-management and photo-sharing startup Mayer ran for roughly six years and about $20 million in funding. Sunshine struggled with traction and privacy questions, and rather than keep pushing it, Mayer wound it down and rolled its team and assets into a fresh company. Its investors, in a graceful bit of housekeeping, received 10% of Dazzle.
So the twelve people now working on Dazzle are, largely, the people who were working on Sunshine. This is either a red flag or a feature, depending on how you feel about second acts. The optimistic read is that a senior team that has failed together, learned, and chosen to try again is a real asset. The skeptical read is that consumer software is brutally hard and reputation is not a product. Both can be true. That is what makes it interesting.
With Dazzle, our goal is to make AI feel simple, helping people get more done and delighting them.