Foundation: 2023
Headquarters: Hangzhou, China
Founder: Liang Wenfeng
Executive Summary: DeepSeek was founded in May 2023 by Liang Wenfeng, who previously established High-Flyer, a quantitative hedge fund in Hangzhou, China. Liang, an AI enthusiast with a background in computer science from Zhejiang University, began his entrepreneurial journey with High-Flyer in 2015, focusing on AI-driven trading strategies. By 2021, High-Flyer was exclusively using AI for its trading, amassing over 10,000 Nvidia A100 GPUs before US export restrictions on AI chips to China were imposed.
The pivot to DeepSeek came from a desire to delve into Artificial General Intelligence (AGI) research, separate from High-Flyer’s financial operations. This move was catalyzed by the global interest in AI following the advent of models like ChatGPT. DeepSeek was initially part of High-Flyer before becoming its own entity in May 2023.
DeepSeek’s approach has been distinct, focusing on open-source AI models and prioritizing innovation over immediate commercialization. Its first significant release was DeepSeek Coder in November 2023, followed by DeepSeek LLM in November of the same year. The company’s breakthrough came with DeepSeek-V2 in May 2024, which not only showcased strong performance but also initiated a price war in China’s AI sector due to its cost-effectiveness.
By December 2024, DeepSeek-V3 was launched, trained with significantly fewer resources than its peers, yet matching top-tier performance. This model, along with subsequent releases like DeepSeek-R1 in January 2025, has positioned DeepSeek as a key player in the global AI landscape, challenging established tech giants and marking a notable moment in AI development.