DeepSeek facts for kids
 |
|
|
Native name
|
杭州深度求索人工智能基础技术研究有限公司
|
|---|---|
| Privately held company | |
| Industry | Information technology Artificial intelligence |
| Founded | 17 July 2023 |
| Founder |
|
| Headquarters | Hangzhou, Zhejiang, China |
|
Key people
|
|
| Owner | High-Flyer |
|
Number of employees
|
Under 200 |
Hangzhou DeepSeek Artificial Intelligence Basic Technology Research Co., Ltd. is a Chinese company that creates artificial intelligence (AI) programs. It is also known by its shorter name, DeepSeek. This company develops special AI programs called large language models (LLMs). These models are like super-smart computer brains that can understand and create human language.
DeepSeek is based in Hangzhou, Zhejiang, China. It is owned and funded by a Chinese investment company called High-Flyer. DeepSeek was started in July 2023 by Liang Wenfeng. He is also the CEO (Chief Executive Officer) for both DeepSeek and High-Flyer. In January 2025, DeepSeek launched its own AI chatbot, also called DeepSeek, along with its main AI model, DeepSeek-R1.
Contents
What Makes DeepSeek Special?
DeepSeek-R1 is a very powerful AI model. It works as well as other top AI models, like OpenAI's GPT-4o. But here's the amazing part: DeepSeek-R1 costs much less to train!
Lower Training Costs
Training a large language model means teaching it using huge amounts of data. This usually costs a lot of money and computer power. DeepSeek says it trained its V3 model for only about US$6 million. This is much less than the $100 million it cost OpenAI to train GPT-4 in 2023. DeepSeek also used about ten times less computing power than Meta used for its similar model, Llama 3.1.
This ability to create powerful AI at a low cost has surprised many people in the AI world. Some have even said DeepSeek is "upending AI." This means it is changing the way things are done in the AI industry.
Open-Weight Models
DeepSeek's models are "open weight." This means that while you can use and change parts of their AI models, you don't have complete freedom to change everything. It's a bit different from truly "open-source" software, where all the code is completely free to use and change.
The company hires smart AI researchers from top Chinese universities. They also hire people from different fields, not just computer science. This helps their AI models learn a wider range of knowledge and skills.
Impact on the AI Market
The low cost of training and running DeepSeek's AI models is a big deal. This breakthrough happened partly because Chinese companies have limited access to special computer chips made by Nvidia. These chips are very important for AI and were restricted by the US government due to trade disagreements.
DeepSeek's success in making AI cheaper and more efficient has sent "shockwaves" through the market. It has challenged the leading AI companies like Nvidia. For example, Nvidia's stock market value dropped by a huge amount after DeepSeek's news came out.
DeepSeek's Influence in China
DeepSeek's models offer great performance at a very low price. This has started a "price war" among AI companies in China. It means other big Chinese tech companies, like ByteDance, Tencent, Baidu, and Alibaba, have also lowered the prices of their AI models.
DeepSeek has even been called the "Pinduoduo of AI." Pinduoduo is a Chinese shopping app known for its very low prices. Even with its low prices, DeepSeek has been able to make a profit, unlike some of its rivals who are losing money.
See also
- DeepSeek (chatbot)
- Artificial intelligence industry in China
- OpenAI
