Partnering with Preview: Lights, Inference, Action
Sequoia By Sonya Huang and James Flynn ● Covered by 2 sources
Preview wants to be the AI video tool teams actually work in, not just another generator. Sequoia says it could become the place where the real video work, and the inference spend, lives.
Based on reporting by Sequoia, By Sonya Huang and James Flynn — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Developers have already had their AI-native moment. Tools like Codex, Claude, and Cursor turned coding into a different kind of motion: faster, more fluid, and frankly hard to compare with the old IDE grind. Video, by contrast, is still wrestling with a workflow built around huge files, uploads, storage, and editing systems that were never designed for this era.
That mismatch is the opening Sequoia is betting on. In its view, generative video has reached the point where the model can act like a new camera. The hard part is no longer just making the pixels move or throwing a $100M production at the problem. The bottleneck has shifted to creativity and inference, and the people making films, TV series, and commercials still need a proper AI-native workflow to keep up.
Preview is the company Sequoia thinks can supply it. The startup’s pitch is straightforward enough: put a video timeline and an infinite canvas for ideation into one collaborative workspace, let teams generate with any model, and keep track of every character, location, and prop with a full record of how each asset was made. That matters because the current setup is messy. Generation, review, and production tracking are split across different apps, and finished takes can lose their source trail before legal clearance is even sorted.
The founders are not newcomers guessing from the outside. Stefan Fejes and Veljko Tornjanski worked together at Veed, then at Vizcom, where Veljko was the first hire and became Head of AI while Stefan led Product. Sequoia says they understand the weird corners of video, move quickly, and ship polished work. More than 100 studios are already using Preview, including agencies making commercials for Fortune 500 brands, AI-native studios, and Hollywood feature filmmakers.
The bigger bet is simple and pretty obvious once you say it out loud: value will land where the work actually happens. If video creation is about to send a lot more inference through the pipeline, the winning product won’t be the prettiest generator. It’ll be the place teams open in the morning and close at night, because that’s where the decisions get made and the paperwork survives.
My take — AI-written commentary, not fact-checked reporting
This is the right kind of AI bet: boring infrastructure with teeth, not another fireworks demo. The industry keeps pretending the model is the product, when the real money usually goes to the workflow that traps the work, the history, and the bill. Funny how often “creative revolution” ends up looking like better project management.
Read more about this at: Sequoia
