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.
免責聲明：投資有風險，本文並非投資建議，以上內容不應被視為任何金融產品的購買或出售要約、建議或邀請，作者或其他用戶的任何相關討論、評論或帖子也不應被視為此類內容。本文僅供一般參考，不考慮您的個人投資目標、財務狀況或需求。TTM對信息的準確性和完整性不承擔任何責任或保證，投資者應自行研究並在投資前尋求專業建議。