Sources
- TechCrunch AI
See it in action
Browse the models and styles behind stories like this one — free account, instant gallery.

Browse the models and styles behind stories like this one — free account, instant gallery.
Snap is spinning off its internal AI video team into a new independent company called Dotmo, with current Snap employees leaving to pursue AI video development full-time — a direct consequence of the prohibitive costs of running frontier video AI inside a social media company.
AI video generation is expensive in a way that image generation simply is not. A single high-quality video clip can require hundreds of times the compute of a comparable still image, and the infrastructure costs compound fast at scale. Snap, which has been experimenting with generative AI features across its platform, apparently found that level of sustained investment difficult to justify inside a consumer social app where the core product is not video generation.
According to TechCrunch, Dotmo will be composed of current Snap staff who are leaving to focus on AI video development — meaning this is not a layoff dressed up as a spinout, but a deliberate structural separation. The team takes its expertise with it, and Snap sheds the ongoing cost burden.
This pattern is becoming familiar. The infrastructure required to train and serve competitive video models — think the kind of quality that rivals Runway, Kling, or Sora — demands dedicated capital allocation that is hard to justify as a line item inside a company whose revenue depends on advertising, not generative AI subscriptions.
For AI-art creators and video artists, a standalone Dotmo is potentially more interesting than an embedded Snap feature. Internal AI video teams at consumer platforms tend to optimize for what serves the parent product — short clips, filters, story formats — rather than the longer, higher-fidelity, prompt-driven outputs that serious creators actually want.
An independent company has different incentives. Dotmo will need to build something people pay for directly, which typically pushes teams toward quality, control, and workflow integration rather than virality. Whether Dotmo pursues a developer API, a standalone generation tool, or something aimed at professional video pipelines is not yet public — but the structural separation at least opens those possibilities.
The cost angle is worth watching closely. As reported by TechCrunch, pricing pressure is already reshaping the AI video market; Avataar AI recently launched video generation at $0.005 per second, setting a new low-end benchmark. Dotmo enters a competitive field where the cost-per-second race is already underway and well-capitalized players like Runway and Google's Veo are setting the quality ceiling.
Creators evaluating AI video generation tools right now are navigating a market in flux — new entrants, aggressive pricing, and frequent capability jumps. Dotmo adds another name to watch, though it will likely be months before any public product surfaces.
This is not Snap's first spinout of an internal unit. The company has a history of incubating technology that proves expensive or strategically tangential and then separating it. The Dotmo move fits that pattern and suggests Snap is narrowing its AI focus to what directly serves Snapchat — personalization, lenses, and on-platform features — rather than building foundational video generation infrastructure.
For creators, the practical takeaway is straightforward: the AI video tools most worth tracking are now being built by dedicated companies with video generation as their entire business, not as a feature inside a social app. Dotmo, once it has a product, will compete on those terms.