OpenAI's Sora Dazzles: How Can Chinese Companies Keep Up With AI That Creates Videos From Text Instructions? (AI Translation)
Listen to the full version
- Chinese AI startups AiShi Technology and Shengshu Technology announced new funding rounds to develop video generation models aiming to match OpenAI's Sora, amidst a cautious approach from major Chinese tech companies due to technological gaps and resource constraints.
- Sora, a significant leap in AI video generation by OpenAI, generates videos from text prompts, showcasing advanced understanding and realism. Despite its potential, the high computational cost and unclear commercial applications make widespread adoption challenging.
- While Chinese firms express ambitions to catch up with or surpass Sora, challenges include the high demand for computational power and the lack of foundational models like LLaMA in language AI. The landscape suggests a cautious investment environment compared to the initial enthusiasm for ChatGPT.

文|财新周刊 杜知航 关聪
By Caixin Weekly's Du Zhihang, Guan Cong
“爱诗科技将集中人力和资源,在三到六个月内赶超Sora目前水平。”
"AiShi Technology will concentrate manpower and resources to surpass the current level of Sora within three to six months."
“生数科技今年年内能达到OpenAI的Sora目前版本的效果。”
"Shengshu Technology will be able to achieve the current version of OpenAI's Sora within this year."
美国公司OpenAI发布人工智能(AI)文生视频大模型产品Sora不到一个月,3月11日、12日,中国的两家AI创业公司爱诗科技和生数科技相继宣布新一轮融资,目标直追Sora。
Less than a month after the U.S. company OpenAI released its artificial intelligence (AI) video generation large model product, Sora, on March 11 and 12, two Chinese AI startups, Aishi Technology and Shengshu Technology, announced new rounds of financing with the goal of catching up with Sora.
