AI editing wars: money shifts from editors toward compute
Several outlets republished the same Dingjiao One feature on the AI video-editing market, which opens with an editor's test letting Claude Opus 5.5 make a short video on its own: it called Fish Audio to clone his voice and HeyGen to drive a digital avatar, then analyzed two reference videos frame by frame and coded the motion effects and sound design, with the editor saying no keyframe was made by hand. The piece describes Jianying's AI workspace Hub, smart editing assistant and AI Ultra membership (199 yuan a month with 22,100 AI credits), Adobe adding agent capabilities to Premiere, and AI-native players like Descript and OpusClip plus video-model companies such as Runway and Kling pushing into fuller production pipelines. The author argues a new pricing model based on credits and usage is emerging, while software vendors are squeezed between rising upstream model and compute costs and users pushing prices down; the efficiency and cost-saving figures cited come from vendor demos or customer cases rather than independent verification.
It signals that AI editing is shifting profits from software subscriptions and editing labor toward models and compute, so toolmakers without distribution or their own models may not convert growth into margin.