Snap Spins Off Its Generative AI Video Team Into Dotmo: Deal Structure, Reasons and What to Watch
Snap is spinning off a generative AI video team into Dotmo, a company building AI models for interactive gaming. How the licence-and-equity deal works, Bobby Murphy's role, and why AI spin-offs are used to manage cost.
Snap, the company behind Snapchat, is spinning off an internal generative AI video team into a separate company called Dotmo. The new company will develop AI models that create interactive gaming experiences, and Snap cited the high cost of doing that work internally as one of the reasons for the move, according to TechCrunch. This article explains how the deal is structured, why companies use spin-offs for expensive AI projects, and what to watch next.
What Dotmo is
Dotmo’s focus is AI models for interactive gaming and entertainment, not features for the Snapchat app. Key points reported by TechCrunch:
- Technology licence. Snap will give Dotmo a licence to adapt Snap’s technology for gaming and interactive entertainment platforms.
- Team. The initial Dotmo team is made up of current Snap employees who are leaving Snap to start the new company.
- Equity. In exchange for the talent and the licence, Snap receives a large equity stake in Dotmo.
- Funding. Snap is not funding Dotmo directly. Dotmo may seek outside funding later.
- Relationship. A Snap representative said Dotmo’s work is not part of Snap’s current core business priorities, but the two could partner in the future if the fit is right.
Bobby Murphy’s role
Snap co-founder and chief technology officer Bobby Murphy is Dotmo’s lead investor and will hold a significant personal stake. He remains at Snap full-time as CTO and continues to lead Snap’s generative AI research and development (TechCrunch; Indian Express).
This creates an unusual arrangement: the executive leading Snap’s AI research is also personally invested in a company built on licensed Snap technology. Arrangements like this are typically governed by board-approved conflict-of-interest policies; how Snap manages it, for example who negotiates future deals between the companies, is something investors are likely to watch.
Why spin off instead of keeping the team
Snap’s stated reason is cost. Training and serving generative video models requires large amounts of compute, and turning research into products, in this case games, adds further spending without a guaranteed return. A spin-off changes who carries those costs:
- The parent’s books. Losses and capital needs move off Snap’s income statement, while Snap keeps upside through its equity stake.
- Access to capital. A standalone startup can raise money from investors interested specifically in AI gaming, who may not want exposure to a social media company.
- Focus. The team can build products that do not have to fit Snapchat’s priorities or user metrics.
- Recruiting. Startup equity can be a stronger incentive for specialised researchers than shares in a large public company.
The trade-off is control. Snap gives up day-to-day direction of the team and shares future value with other investors.
Context: Snap’s other 2026 restructuring
TechCrunch notes this is Snap’s second major spin-off effort of 2026. Earlier in the year, Snap moved its smart-glasses work into a separate operation focused on its Specs line; the recently unveiled Specs drew attention for a price of around $2,200. Snap also went through layoffs earlier in the year. Taken together, the moves suggest a company narrowing its core while keeping options on longer-term bets.
Common ways to separate a costly team
Companies with expensive research teams usually choose between keeping the team in-house, moving it into a wholly owned subsidiary, spinning it out with outside investors, or selling it. Dotmo follows the third route: the parent contributes a technology licence and people in exchange for equity, does not fund the company directly, and leaves it to raise outside capital later. That keeps upside for the parent while moving development cost and risk off its books; the price is less control and the need to negotiate any future partnership at arm’s length.
What it means for generative AI video
Three broader points follow from the deal:
- Generative video is expensive to develop. Even a large consumer platform judged that building AI-driven interactive experiences in-house was costly enough to restructure around.
- Gaming is a target market. Interactive, AI-generated worlds are an area several AI labs are exploring, and Dotmo positions Snap’s research for that market rather than for social features.
- Licence-plus-equity structures let a parent keep intellectual property and upside while separating risk. Watch whether other companies use similar structures for costly AI units.
What to watch
- Whether Dotmo raises outside funding, and from whom.
- The first products or demos Dotmo releases, and on which platforms.
- Whether Snap and Dotmo announce a commercial partnership, and how it is structured given Murphy’s dual role.
- How Snap’s own AI features evolve without this team.
Sources
Reader Dispatches
No reader responses yet. Share your technical perspective below.
Related Posts
Epic Games Shows How Generative AI Fits Into Fortnite's Concept Art Process
In a June 2026 video, Epic Games showed how Fortnite artists start from hand-drawn designs and 3D block-outs, then use its internal GenMedia Bridge tool with Nano Banana and GPT Image for rendering and variations. Here is what the video shows, what Epic said, and what remains unanswered.
Epic Games Confirms Generative AI in Fortnite Art Pipeline — and the Gaming Industry Is Watching
Epic Games breaks the silence on AI-generated assets in Fortnite, revealing a hybrid human-AI workflow for character skin and concept art design. The disclosure comes amid a wave of generative AI controversies across major game studios.
Krea 2 Goes Open Source: What the RAW and Turbo Split Means for Image Generation
Krea AI has open-sourced Krea 2, its aesthetic-first image model, as two checkpoints: RAW for LoRA training and Turbo for 8-step inference. Here's how the architecture, the license, and the day-one ecosystem stack up.