Orchestrating AI Video for Marketers
Alex Mashrabov, CEO of Higgsfield, on orchestrating AI video models
The failed mobile retention was the clearest signal that Higgsfield was not a mass consumer selfie app, it was a professional production tool hiding inside one. The users who stayed were VFX artists and marketers who needed desktop level control, repeatable workflows, and better outputs for ads and storyboards. That repositioning turned fast but shallow consumer demand into a much stronger professional wedge, and the business accelerated from there.
-
The mobile app proved there was broad curiosity around AI video, but the durable need came from users making commercial work. Higgsfield later found fit with agencies and e-commerce brands creating video ads, product demos, and social content, then grew from $11M ARR in May 2025 to $100M in November 2025 and $230M in January 2026.
-
Mobile was the wrong surface because pro users need fine control over prompts, shots, effects, and iteration. Higgsfield responded by moving to desktop, where it could package multiple models into preset workflows and cinematic effect galleries, which is much closer to how a marketer or editor actually produces ad creative.
-
This also defines the competitive lane. Runway is more vertically integrated around filmmaker grade tooling and proprietary video systems, while OpenArt has leaned toward simpler creator and SMB workflows. Higgsfield sits between them as a marketer focused orchestrator, bundling models and hiding the complexity behind templates and auto selection.
The next step is turning from a creation tool into a full marketing workflow, from ideation and generation to publishing and measurement. If Higgsfield keeps compounding around professional users instead of chasing broad consumer novelty, it can become the system marketers use every day to make, test, and scale AI native video ads.