Market research agency Factorial Funds recently released a report stating that OpenAI's deployment of the text-to-video model Sora would require 720,000 Nvidia H100 AI accelerators during the peak period. Based on the cost of $30,000 per Nvidia H100 AI accelerator card, 720,000 would cost $21.6 billion (currently approximately RMB 156.168 billion). However, this is only the cost of an H100 accelerator card. If you want to run it, it will consume a lot of electricity. Each H100 consumes 700W of power, and 720,000 is 504 million watts. Factorial Funds estimates that training Sora for 1 month would require 4200 to 10,500 Nvidia H100 AI accelerators, and 12 minutes to render a 1-minute video if using a single H100 accelerator card.
机构预估：部署Sora需要72万片英伟达加速卡，价值216亿美元
Agency estimates: deploying Sora requires 720,000 Nvidia accelerator cards worth $21.6 billion
The translation is provided by third-party software.
The above content is for informational or educational purposes only and does not constitute any investment advice related to Futu. Although we strive to ensure the truthfulness, accuracy, and originality of all such content, we cannot guarantee it.