// news · multimodal2026-06-16source: openai / magichour / opus

OpenAI discontinues Sora web and app experiences with API winding down September 24 — rare strategic retreat cedes consumer video generation to Veo, Seedance, and Kling

Sora's web and app shutdown April 26 with API end-of-life September 24 marks OpenAI's exit from consumer video generation. The withdrawal cedes the category to Google Veo, ByteDance Seedance, and Kling — a rare strategic retreat for OpenAI in a frontier modality, and a structural realignment of the video-gen competitive landscape.

The substantive piece is the strategic-retreat asymmetry. OpenAI has rarely exited frontier modalities once entered; Sora's discontinuation marks the first such retreat in the company's product history. The September API sunset gives enterprise customers ~3 months to migrate; the migration paths route to Veo (Google ecosystem), Seedance (Chinese-vendor stack), or Kling (Chinese-vendor stack). H2 2026 video-gen procurement defaults will be set by these migration decisions rather than by greenfield evaluations.

The competitive read against Seedance 2.0's leaderboard displacement of Kling is that two of three Sora migration destinations are Chinese-vendor stacks. The H2 2026 video-gen market segmentation lands with Google Veo holding the US-vendor + audio-sync niche, Chinese vendors holding the capability-frontier + multi-shot tiers, and Runway holding the workflow-anchored marketer segment. Sora's exit removes the only meaningful US-vendor standalone-platform alternative to Google Veo.

See our analysis →

Magic Hour — AI Video Model Release Tracker 2026 → · Opus Pro — Gemini Omni Multimodal Video Model Explained → · Build MVP Fast — Best LLMs 2026: Video Generation AI →