What is DeepSeek and Why is it Disrupting the AI Sector?

What is DeepSeek and For what reason is it Upsetting the simulated intelligence Area?

BEIJING Jan 27 (Reuters) Chinese startup DeepSeek has shaken up the global AI industry by launching advanced AI models that rival leading U.S. counterparts at a fraction of the cost. With its latest model DeepSeek V3 the company has gained international attention claiming to match or outperform top tier AI systems while significantly reducing training expenses.

One of the Biggest revelations came when DeepSeek disclosed that training DeepSeek V3 required less than $6 million worth of computing power using Nvidia H800 chips. This achievement has cast doubt on the billion-dollar AI investments made by U.S. tech giants and has even impacted the stock prices of major players, including Nvidia.

DeepSeek AI assistant powered by

DeepSeek-V3, has surpassed ChatGPT to become the top-rated free app on Apple’s U.S. App Store, further solidifying its position in the AI landscape.

Why is DeepSeek Making Waves?

Since OpenAI launched ChatGPT in late 2022, Chinese tech companies have rushed to develop their own AI-powered chatbots. However, the initial Chinese alternatives, such as Baidu’s ERNIE Bot, failed to match the capabilities of Western AI models, reinforcing concerns about China’s AI lag.

DeepSeek, however, has flipped this narrative. Its models, DeepSeek-V3 and DeepSeek-R1, have gained recognition from Silicon Valley executives and U.S. tech engineers, who claim they are on par with OpenAI and Meta’s most advanced models.

Scale AI CEO Alexandr Wang stated in an interview with CNBC that DeepSeek allegedly possesses 50,000 Nvidia H100 chips a controversial claim, as U.S. export control prohibit the sale of such advanced AI chip to Chinese companies. DeepSeek has not commented on these allegation.

Despite DeepSeek claims that V3 was trained using just $5.58 million in computing power analysts from Bernstein argue that the actual training cost could be much higher. Similarly the training costs for DeepSeek R1 remain undisclosed.

Who is Behind DeepSeek

DeepSeek is a Hangzhou-based startup controlled by Liang We Feng co founder of the quantitative hedge fund High Flyer.

In March 2023 High-Flyer announced its shift beyond financial trading stating it would focus on developing Artificial General Intelligence (AGI) autonomous systems that surpas humans in most economically valuable task. DeepSeek was established later that year.

While the exact investment from High-Flyer in DeepSeek is unknown, corporate records indicate that High-Flyer shares an office building with DeepSeek and holds patents related to AI chip clusters. In July 2022, High-Flyer’s AI division revealed it owned and operated 10,000 Nvidia A100 chips further signaling its deep involvement in AI research.

How Does Beijing View DeepSeek?

DeepSeek’s rapid success has not gone unnoticed by China’s top political leaders. On January 20, the day DeepSeek R1 was released, company founder Liang We Feng attended a closed-door symposium with Chinese Premier Li Qiang, as reported by state news agency Xinhua.

Liang cooperation recommend that DeepSeek could assume a part in Beijing methodology to counter U.S. send out limitations and accomplish mechanically independence in man made intelligence and semiconductor ventures. A comparable gathering last year was gone to by Baidu President Robin Li building up the significance of man-made intelligence improvement in China identity plan.

With its problematically computer based intelligence leap forwards DeepSeek is arising as a vital participant in the world wide man made intelligence race one that might actually Challenge U.S. predominance in the field.