据市场调查机构Factorial Funds最新报告显示,人工智能研究机构OpenAI在部署其先进的文本生成视频模型Sora时,将面临巨大的硬件需求。该机构预估,在模型运行的峰值阶段,OpenAI可能需要高达72万张英伟达的顶级H100 AI加速卡,这将产生总计216亿美元的硬件成本。
Factorial Funds在报告中指出,为了训练Sora模型,OpenAI在一个月内可能消耗4200至10500片H100加速卡,显示出Sora对高性能计算资源的惊人胃口。值得注意的是,即使使用单片H100加速卡,生成短短1分钟的视频也需要大约12分钟的渲染时间,这揭示了文本转视频技术在当前阶段的复杂性和计算密集度。
这一估算反映了AI研发领域对硬件资源的庞大需求,同时也揭示了OpenAI在推进创新技术过程中可能面临的经济和技术挑战。随着AI技术的不断发展,对高性能计算设备的依赖度只会增加,这不仅影响到研究成本,也对整个AI行业的可持续发展提出了新的问题。OpenAI尚未就Factorial Funds的报告发表官方评论,但这一估算无疑为业界提供了一个关于尖端AI项目规模和成本的参考。
英语如下:
News Title: “Staggering Cost of Deploying OpenAI’s Text-to-Video Model Sora: 720,000 NVIDIA H100 Cards, Valued at $216 Billion”
Keywords: OpenAI, Sora model, NVIDIA H100
News Content: According to the latest report from market research firm Factorial Funds, AI research institute OpenAI is set to face substantial hardware requirements in deploying its advanced text-to-video model, Sora. The report estimates that during peak operation, OpenAI might require as many as 720,000 top-of-the-line NVIDIA H100 AI accelerator cards, translating to a hardware cost of approximately $216 billion.
Factorial Funds highlights in its report that for training the Sora model, OpenAI could consume between 4,200 and 10,500 H100 cards in a single month, demonstrating the model’s immense appetite for high-performance computing resources. Notably, even with a single H100 card, generating a 1-minute video reportedly takes about 12 minutes of rendering time, underscoring the complexity and computational intensity of text-to-video technology at its current stage.
This estimation reflects the massive demand for hardware resources in the AI research field and the economic and technological challenges OpenAI might encounter in advancing innovative technologies. As AI technology continues to evolve, reliance on high-performance computing devices will only increase, impacting research costs and posing new questions about the sustainability of the AI industry. OpenAI has not yet commented officially on Factorial Funds’ report, but this estimation undoubtedly provides an industry reference for the scale and cost of cutting-edge AI projects.
【来源】https://www.ithome.com/0/758/413.htm
Views: 4