Snap has a straightforward problem. It has been building things it cannot afford. The company posted a $460 million net loss in 2025 on nearly $6 billion in revenue. That is not a startup burning cash to grow fast. That is a mature platform that has never figured out how to be consistently profitable.
So in 2026, Snap is doing something it has avoided for years: letting things go. In April, it cut around 1,000 people, representing about 16% of its full-time workforce. The stated goal was to reduce annual costs by more than $500 million. Now it is spinning out its AI video research team into a new company called Dotmo.
Dotmo will focus on building AI tools that generate interactive gaming experiences. The people building it are current Snap employees who are leaving to run the new venture. Snap is giving them a license to use its existing technology and, in return, gets a large ownership stake in the new company. Snap's co-founder and chief technology officer Bobby Murphy is putting in his own personal money as the lead backer, while continuing in his full-time role at Snap.
This structure tells you something. Snap clearly thinks the underlying technology has potential: otherwise Murphy would not be investing his own money. But the cost of doing this work inside a public company trying to reach profitability is too high to justify right now. Spinning it out solves the problem cleanly. The cost leaves the balance sheet. The upside stays on it, through equity.
Building the kind of AI models that can generate real-time interactive video is extraordinarily expensive. Training a single competitive AI model can cost hundreds of millions of dollars in computing power. Snap, which is still not reliably profitable, cannot carry that bill internally while also trying to convince investors it is on a path to sustainable earnings.
This is Snap's second spinoff of 2026. The Specs smart glasses division was pushed out into its own company in January, partly because an activist investor was pushing Snap to cut or sell it. The Specs launch earlier this week drew attention for its $2,195 price tag, steep enough to put it out of reach for most everyday buyers, and Snap's stock dropped on the news.
Dotmo is a different kind of exit. Unlike Specs, which is central to Snap's long-term hardware ambitions, Dotmo is working on things Snap has publicly said are not part of its core business right now. Gaming and interactive AI experiences are real markets, but they are not where Snap makes money today. Advertising on Snapchat is.
For anyone watching how companies handle expensive AI research, this model is worth noting. A parent company funds early work, a team develops something promising but costly, and then the team is moved out with a technology license and outside funding. The parent keeps equity and loses the operating expense. It is a structure that more traditional companies may look at as AI costs keep climbing.
Dotmo may also seek outside investment separately from Murphy's stake. If it does, Snap's equity position could become worth real money. If it does not work out, Snap's exposure is limited to the technology license and the people it already spent money training.