How DeepSeek, Chinese AI is buzzing in tech and shaking up ChatGPT, Nvidia

The tech industry is abuzz over Chinese startup DeepSeek’s latest artificial intelligence (AI) models, which they claim rival or exceed the best US models in quality while being considerably cheaper. The company gained international recognition after disclosing that training its DeepSeek-V3 model required less than $6 million with Nvidia H800 chips. DeepSeek’s AI Assistant, driven by DeepSeek-V3, has now achieved the status of the top free app on the Apple App Store in the US, outpacing ChatGPT. This has sparked questions regarding the motivation behind US tech firms pouring billions into AI, as shares of major companies like Nvidia have experienced a downturn.

DeepSeek is an AI platform that offers sophisticated artificial intelligence models, especially for functions like natural language processing, data evaluation, and multiple machine learning uses. Models like DeepSeek-V3 and DeepSeek-R1 are engineered to be extremely efficient, economical, and adept at managing complex tasks with great precision.

The launch of OpenAI’s ChatGPT in late 2022 sparked a competition among Chinese technology companies to create their own AI-driven chatbots. Nevertheless, following the debut of Baidu’s initial Chinese counterpart to ChatGPT, there was significant discontent in China because of the disparity in AI capabilities between American and Chinese companies. DeepSeek has altered this story with its models, DeepSeek-V3 and DeepSeek-R1, which have received accolades from Silicon Valley leaders and US technology engineers. The Chinese startup asserts that these models rival the most advanced products from OpenAI and Meta.

Additionally, they are significantly more economical. According to a post on DeepSeek’s official WeChat account, the recently launched DeepSeek-R1 is 20 to 50 times less expensive to utilize than OpenAI’s GPT-3 model, varying by the task.  According to Chinese corporate records, DeepSeek, a startup located in Hangzhou, is managed by Liang Wenfeng, who is a co-founder of the quantitative hedge fund High-Flyer.

In March 2023, Liang’s fund revealed a change in direction on its official WeChat account, transitioning from trading to establishing a “new and independent research group” centered on Artificial General Intelligence (AGI). DeepSeek was established later that same year as a result of this new approach. OpenAI, the developers of ChatGPT, describes AGI as self-sufficient systems that can surpass human performance in the majority of economically important tasks.

The extent of High-Flyer’s investment in DeepSeek remains uncertain. The two firms occupy the same office area, and High-Flyer possesses patents for chip clusters utilized in AI model training, based on corporate documentation. In July 2022, the fund’s AI sector announced on WeChat that it manages a cluster of 10,000 A100 chips.

The success of DeepSeek has attracted the interest of leading political figures in China. On January 20, the day of DeepSeek-R1’s launch, founder Liang Wenfeng attended a private symposium for business leaders and experts organized by Chinese Premier Li Qiang, according to state news agency Xinhua. Liang’s participation in the event indicates that DeepSeek’s successes may coincide with Beijing’s aim of surmounting U.S. export limitations and attaining independence in crucial areas such as AI. A comparable symposium last year featured Baidu CEO Robin Li.