OpenAI’s Sora generates 5 minutes of videos produced per NVIDIA H100 per hour, equivalent to 120 minutes of videos per H100 per day, according to estimates from Factorial funds. 

The report further adds that an estimate of approximately 89,000 NVIDIA H100 GPUs needed to support the creator community on TikTok and YouTube. Combining the AI-generated video production from TikTok and YouTube yields a total of 10.7 million minutes of videos produced daily by AI. 

However, taking into account factors such as realistic utilisation, peak demand and busy traffic, the estimated number of Nvidia H100 GPUs needed at peak demand is approximately 720,000, significantly higher than the initial calculation based on simplified assumptions. 

Creators are likely to generate multiple candidate videos before selecting the best one, leading to an average of two candidates per uploaded video. This factor also doubles the GPU requirements.

In a recent interview with the Wall Street Journal, CTO Mira Murati shared that OpenAI will make Sora publicly accessible later this year. When Sora was launched earlier in February, users greatly appreciated its hyper-realistic videos, many calling it the “ChatGPT moment for video”

The model, showcased in February, generates realistic scenes from text prompts and will soon be open for public use. The initial rollout will primarily target visual artists and filmmakers. Murati also disclosed plans to incorporate sound and editing flexibility into Sora-generated videos. 

OpenAI is pitching Sora to Hollywood. The ChatGPT creator has scheduled meetings in Los Angeles next week with Hollywood studios, media executives and talent agencies to form partnerships in the entertainment industry and encourage filmmakers to integrate its new AI video generator into their work, reported Bloomberg.