China’s Deepseek Emerges As A Game-Change In AI

By Editor Team

Both the debates around Artificial Intelligence (AI) and the popularity of AI are increasing rapidly across the world. The landscape of AI is constantly changing and the growing popularity of AI is leading to the emergence of new players in the market. In this regard, A new AI model from China called DeepSeek is gaining attention in the global tech sector in a short span of time and posing an intense competitive threat to the other popular AI models like Gemini, ChatGPT and Claude in terms of performance.

About DeepSeek

DeepSeek is an advanced Chinese AI chatbot which was developed by an independent  Hangzhou-based research lab or startup company of the same name. The startup company was established in 2023 by a Chinese entrepreneur, Liang Wenfeng, belonging to the age group of 40 years old. Liang Wenfeng, the CEO of the company, is operating the research lab under the umbrella of a leading Chinese quantitative hedge fund called High-Flyer.

Through a post on X (Formerly Twitter) on 15th January 2025, the company introduced their official app powered by an advanced open-source AI model, V3, which is available on the Google Play store and App Store. Since its launch, the app has come to the top in the download charts on iPhone’s App Store in multiple countries like the US, China and the UK, surpassing OpenAI’s ChatGPT app.

A fraction of the cost of the rivals has been used for the development of the new Chinese chatbot which makes the AI dominance of America in the future questionable. By releasing a research paper, the development team has said that they had spent only two months and less than six million dollars on computing power to train the open-source, free large language model V3. The reduced capability chips H800s from Nvidia are used to develop this AI model.

On January 20, 2025, DeepSeek also introduced its latest release, R1, which is found to be showcasing similar performance or even more in some third-party tastes compared to the rival industry leader OpenAI’s o1 model. However, RI is a fully open-source reasoning model and comparatively more cost-efficient and offers unlimited free usage whereas OpenAI’s o1 model which charges 200 dollars per month from the users. The R1 model has achieved high performance without the high investment in comparison to the other technologies providing such performance.

Impact on the US share market

The emergence of DeepSeek AI has become a concern for the US tech giants associated with AI. The US market, which has dominated globally in the context of introducing new AI technology, the stock market of the US has now crashed after the emergence of this Chinese AI. On 27th January 2025, the share value of the leading US chipmaker Nvidia, which is one of the companies that benefited the most in the boom of AI, dropped around 18 per cent with the increasing popularity of the new Chinese AI maker.

As a result, the market value of Nvidia declined by 600 billion dollars, which led to the market value of the company below 3 trillion dollars. In the history of the US stock market, this became the biggest drop in the market value. According to some analysts, the lost market value of Nvidia on 27th January 2025 is higher than the entire market value of Costco, Exxon Mobil or Bank of America.

The share of other prominent AI related organizations, including Microsoft, Alphabet, and Meta, has also sunk in response to the success of DeepSeek. The stock value of Google’s parent company, Alphabet, dropped by more than 4 per cent, while Microsoft faced a 2.14 per cent drop in the market value. According to the statement of the CEO of Microsoft, Satya Nadella, the new AI model from China is super impressive and highly effective. He also mentioned that they should consider the developments and innovations from China quite seriously.

Challenges faced by DeepSeek 

After gaining huge sudden popularity, DeepSeek has been seen to become the victim of a cyberattack. As a result, in a statement, the company said on 27th January that with the aim of ensuring the continuity of their service offering, they are temporarily limiting the new registrations as their services have confronted large-scale malicious attacks. They also added, “Registered users can log in normally. Thank you for your understanding and support.” Earlier in the day, the startup company was also seen to face outages in its website after the skyrocketing popularity of its AI assistant.

DeepSeek also had to face the challenges of semiconductor shortage. The US government imposed strict restrictions on China in the access of most advanced chips, such as Nvidia’s H100s, to prevent them from being involved in AI advancement. Nevertheless, around these restrictions, the company was able to introduce its latest AI models with the involvement of fewer resources.

The AI market is full of competition. Though the tech analysts are found to be agreeing that the newly released R1 model has a similar level of performance to ChatGPT or even shows better performance in certain aspects, they also said that the AI field is rapidly evolving. Earlier this month, the CEO of OpenAI, Sam Altman, revealed that the company is in the process of releasing its most powerful reasoning AI model, o3, within weeks after the consideration of user feedback.

However, recently, through a post on X by Altman, the R1 model has been recognized as an impressive model, particularly because of its pricing strategy. Nevertheless, he also added that they will obviously develop a much better model with higher performances.

Furthermore, DeepSeek is not the only organization from China which is involved in the development of higher performing AI models with limited investment. Kai-Fu Lee, a leading Chinese AI researcher, has said that only 3 million dollars has been invested to train his startup O1.ai. Moreover, by releasing an update to its AI model, the parent company of TikTok, ByteDance, has claimed that its AI model has the potential to show outperformance compared to OpenAI’s o1 in a major benchmark test.

FAQ

1. What is DeepSeek?

This is an advanced Chinese AI chatbot developed by an independent Hangzhou-based research lab or startup company of the same name. The startup was established in 2023 by Chinese entrepreneur Liang Wenfeng.

2. How did China manage to compete with major AI models like ChatGPT?

A fraction of the cost of the rivals has been used for the development of the DeepSeek chatbot which makes the AI dominance of America in the future questionable. By releasing a research paper, the development team has said that they had spent only two months and less than six million dollars on computing power to train the open-source, free large language model V3. The reduced capability chips H800s from Nvidia are used to develop this AI model.

3. What is the R1 model, and how does it compare to OpenAI’s o1 model?

On January 20, 2025, DeepSeek also introduced its latest release, R1, which is found to be showcasing similar performance or even more in some third-party tastes compared to the rival industry leader OpenAI’s o1 model. However, RI is a fully open-source reasoning model and, comparatively more cost-efficient and offers unlimited free usage, whereas OpenAI’s o1 model charges 200 dollars per month from the users.

4. How has DeepSeek impacted the US stock market?

The market value of Nvidia declined by 600 billion dollars. The share of other prominent AI-related organisations including Microsoft, Alphabet, and Meta, has also sunk in response to the success of the new Chinese AI model. 

5. What challenges has DeepSeek faced?

After gaining huge sudden popularity, the company has been seen to become the victim of a cyberattack.

Follow us on

Share This Article
Leave a comment

Leave a Reply

Your email address will not be published. Required fields are marked *

Saturday, Feb 22, 2025