Chinese artificial intelligence startup DeepSeek unveiled a free open-source large language model last week, which it says uses less data versus existing services.
DeepSeek, the new model, is open-sourced, meaning developers can access and modify it for training.
The startup said its performance is comparable to the official version of OpenAI, the developer of ChatGPT.
Founded in 2023 in Hangzhou, China, its chief executive officer Liang Wenfeng previously co-founded hedge fund High-Flyer, which invests in AI-driven quantitative trading.
On Monday, the startup was hit with outages after DeepSeek became the top-rated free app on Apple's App Store in the US, Reuters reported.
DeepSeek uses Nvidia chips to train its systems, which are not banned in China, the Associated Press reported.
Its launch triggered a selloff of AI and tech stocks globally, with Nvidia posting a 17% drop on Monday and losing $600 billion in value. Nvidia recovered on Tuesday, gaining 8.8%.
Chip makers and designers such as AMD, Arm, Broadcom and Micron have also suffered losses Monday, while AI tech behemoths Alphabet, Amazon, Meta Platforms and Microsoft, which all invest heavily in AI, traded lower on Monday.
The Nasdaq Composite fell 3% on Monday before recovering on Tuesday with a 2% gain.
免责声明：投资有风险，本文并非投资建议，以上内容不应被视为任何金融产品的购买或出售要约、建议或邀请，作者或其他用户的任何相关讨论、评论或帖子也不应被视为此类内容。本文仅供一般参考，不考虑您的个人投资目标、财务状况或需求。TTM对信息的准确性和完整性不承担任何责任或保证，投资者应自行研究并在投资前寻求专业建议。