OpenAI sunsets the Sora product as AI video moves from demos into production pipelines
OpenAI discontinued the Sora web and app experiences on 26 April 2026, with the deprecated API to follow on 24 September — even as its Sora 2 model competes at the frontier. The retreat from a standalone product tracks the year's real shift: AI video is moving out of standalone text-to-video demos and into embedded production workflows.
Killing the standalone product while keeping the model is a strategic read of where the value is. Sora 2 remains one of the four frontier video models, but the consumer app around it did not find durable footing — because the demand is increasingly for video generation embedded in professional pipelines, not a novelty destination site. OpenAI is following the value into the workflow.
The broader pattern is that 2026 AI video left the demo phase. The biggest change in the category is not another quality jump but the move into full production workflows — editing suites, avatar platforms, enterprise creative tools — where generation is one step among many rather than the whole product. Standalone text-to-video was the demo; embedded generation is the business.
The consolidation cuts both ways. As standalone consumer apps retreat, the surviving contenders — Seedance, Kling, Veo, and Sora-2-as-a-model — compete to be the generation engine inside other people's tools. The distribution question becomes which pipelines adopt which model, not which app a consumer opens, and that is a quieter, more durable market.
Wikipedia — Sora (text-to-video model) → · Data Science Collective — The 2026 AI video production playbook →