What is DeepSeek and How is it Transforming the AI Industry? – One America News Network
By Eduardo Baptista
January 27, 2025 – 6:17 AM PST
Advertisement
BEIJING (Reuters) – The recent introduction of advanced AI models by the Chinese startup DeepSeek has raised eyebrows in the tech industry, as the company claims its models match or even exceed those of leading American counterparts but at a significantly lower cost.
DeepSeek gained traction after revealing last month that the development of its DeepSeek-V3 relied on under $6 million in computing capacity using Nvidia H800 chips.
Powered by DeepSeek-V3, DeepSeek’s AI Assistant has surpassed ChatGPT to become the highest-rated free application on Apple’s App Store in the U.S.
This achievement has cast doubt on the strategic investments made by various U.S. tech firms, and shares of major players such as Nvidia have experienced declines.
Here are some important points regarding the company that is making waves in the global AI landscape.
WHAT IS CAUSING DEEPSEEK’S DISRUPTION?
The debut of OpenAI’s ChatGPT in late 2022 triggered a rush among Chinese technology companies to develop their own AI-driven chatbots.
However, the first Chinese counterpart released by Baidu (9888.HK) led to widespread dissatisfaction regarding the disparity in AI competency between U.S. and Chinese companies.
DeepSeek has fundamentally altered this narrative with the acclaim received for its models, DeepSeek-V3 and DeepSeek-R1, which it asserts rival OpenAI and Meta’s top technologies.
Moreover, DeepSeek’s models are substantially more affordable. The recently unveiled DeepSeek-R1 is said to cost between 20 to 50 times less than OpenAI’s o1 model, depending on the specific application, as noted on DeepSeek’s official WeChat account.
Yet, there are skeptics questioning DeepSeek’s rapid rise.
During a CNBC interview, Scale AI CEO Alexandr Wang claimed, without providing evidence, that DeepSeek possesses 50,000 Nvidia H100 chips, which he suggested could not be disclosed due to Washington’s export restrictions on advanced AI technology sales to Chinese entities. DeepSeek has yet to respond to this accusation.
Analysts from Bernstein noted on Monday that the actual training expenses for DeepSeek’s V3 model remain unclear and are likely much more than the reported $5.58 million, with no information provided on the training costs for the well-regarded R1 model.
WHO IS THE FORCE BEHIND DEEPSEEK?
DeepSeek, based in Hangzhou, is predominantly controlled by Liang Wenfeng, a co-founder of the quantitative hedge fund High-Flyer, as per Chinese corporate filings.
In March 2023, Liang’s fund announced a strategic shift via its official WeChat account, focusing its resources on establishing an “independent research group to delve into the essence of AGI” (Artificial General Intelligence). DeepSeek was subsequently launched later that same year.
ChatGPT creators, OpenAI, describe AGI as systems capable of outperforming humans in most economically valuable activities.
The investment details from High-Flyer into DeepSeek remain unclear, but it operates from the same building as DeepSeek, and holds patents related to chip clusters utilized for AI model training, according to Chinese records.
High-Flyer’s AI division claimed on its WeChat account in July 2022 that it manages a cluster of 10,000 A100 chips.
WHAT IS BEIJING’S PERSPECTIVE ON DEEPSEEK?
The achievements of DeepSeek have caught the attention of China’s top political figures. On January 20, the same day the DeepSeek-R1 was launched, founder Liang participated in a private symposium for business leaders and experts convened by Chinese Premier Li Qiang, as reported by state news agency Xinhua.
Liang’s involvement in the meeting may signify that DeepSeek’s progress aligns with Beijing’s ambitions to navigate Washington’s export controls and enhance self-sufficiency in critical sectors such as AI.
A similar assembly last year featured Baidu CEO Robin Li.
Reporting by Eduardo Baptista. Editing by Jane Merriman
Advertisements below
