DeepSeek is making similar AI modules like OpenAI and Google, but the big thing is that DeepSeek AI is doing all these at much lower costs in comparison to these tech giants, also within a short period of time of less than two years. Interestingly, DeepSeek’s models are working very effectively, like OpenAI’s ChatGPT, even after spending just around $6 million. This has left tech giants completely surprised.
What is DeepSeek?
DeepSeek is a Chinese software company. After the words of U.S. President Donald Trump,’ this company has gained more popularity. This company works on AI and has made some big achievements in the field of AI worldwide. This company was founded less than two years ago in December 2023, but still, this company has attracted attention worldwide because of its highly advanced and reliable AI modules, which are capable enough to compete with some of the renowned tech giants like OpenAI and Google.
Who is Liang Wenfeng?
Liang Wenfeng was born in 1985 in Zhanjiang, Guangdong, China. He is the founder of DeepSeek. Both of his parents were school teachers in China. During his school days he used to spend much of his time with computers and trying to do something different. He later studied computers and machines at Zhejiang University. Liang received a Bachelor of Engineering in electronic information engineering in the year 2007 and then a Master of Engineering in information and communication engineering in the year 2010.
Later Liang Wenfeng started High-Flyer in the year 2016, simply a type of business that helped people make money with technology. High-Flyer was the first Chinese quantitative hedge fund to raise over 100 billion yuan.
Liang is always thinking of new ideas in field of technology. He believes, “China should create its own future; he believes that China’s AI sector should aim as high as possible. China just can’t keep copying from others forever and should believe in originality and creating something on their own rather than just copying from others.
What Has DeepSeek Done?
What DeepSeek AI has done or achieved is just like a bad nightmare for tech giants like OpenAI and Google. After all, China is again being like typical China by creating something at such a cheap price that the other countries can even think about doing.
DeepSeek’s AI models, including DeepSeek-V3 and DeepSeek-R1, are now among some of the best AI modules in the world. And what’s actually shocking is the cost behind the creation of both of these AI modules. The cost behind the training of DeepSeek-V3 for its computing power is just around $6 million; this will seem like something unbelievable if we compare it with the cost of companies like OpenAI and Google, who spend billions of dollars to do the same thing in AI modules.
The cost difference between DeepSeek’s AI and other AI modules
- The major cost saver for DeepSeek AI is the cost of chips that are being used in their AI modules. For instance, DeepSeek AI used about 2,000 Nvidia H800 chips in their AI system. These chips are not the best chips that are available right now and are not even very advanced chips that would have been used to build their system. But still, their AI modules are working really efficiently even with older and less advanced chips.
However, Scale AI’s CEO has claimed that DeepSeek AI might have access to those advanced chips that are actually restricted by U.S. export controls. But DeepSeek has made it clear, and they say it uses older chips only, including a stockpile of 10,000 Nvidia A100 chips that were purchased much before the restrictions were made. - Not only is the cost to build the system cheaper, but DeepSeek’s models are also very affordable to operate and use. For doing some tasks, DeepSeek-R1 is 20 to 50 times cheaper than OpenAI’s O1 model, and the level of efficiency DeepSeek’s modules are providing to its users at such low costs has made DeepSeek’s AI assistant one of the top-rated free apps on Apple’s App Store in the United States.
How DeepSeek AI is Changing the AI World?
DeepSeek AI’s success over a short period of time has already brought some big changes in the tech industry ruled by a few tech giants. It has been proven that no one should underestimate new and smaller companies when you’re in the tech industry. Because even new and smaller companies can compete with tech giants more smartly and efficiently. This DeepSeek has somehow shaken the confidence of U.S. companies who were taking themselves as only leaders in the race of AI.
When DeepSeek AI announced its success on the 27th of January 2025, many companies’ stocks started falling. Nvidia, a maker of computer parts, lost around $600 billion in the company’s valuation, and even Google and Microsoft saw their stock prices drop.
Why DeepSeek is Important for China?
China badly wants to beat the U.S. in the global race to lead the world. China also wants to lead in creating its own technology, and DeepSeek has started to show the right way to China.
DeepSeek AI’s work is so important that Liang Wenfeng has recently attended a major meeting with China’s government leaders.
Is DeepSeek Winning the AI Race?
It’s really tough to say if DeepSeek AI is winning the race of AI or not. As of now, DeepSeek’s models are actually impressive, but the AI field moves very fast, and things change quickly. Companies like OpenAI and Google are constantly working on their technologies to bring improvements. OpenAI’s CEO, Sam Altman, has acknowledged DeepSeek’s success but promised better models are yet to come from OpenAI’s side.
Still, DeepSeek’s success and its way of working have changed the game a lot. It has shown that small and new companies don’t need billions of dollars or only the latest hardware to create an amazing AI system.
FAQs
- Who founded DeepSeek?
Ans: DeepSeek was founded by Liang Wenfeng in December 2023. - What makes DeepSeek unique?
Ans: DeepSeek’s AI models are cost-effective and powerful in comparison to companies like OpenAI and Google but cost much less to create and use. - How much did DeepSeek spend on creating its AI model?
Ans: DeepSeek AI had spent less than $6 million on computing power to train its DeepSeek-V3 model. - How has DeepSeek affected U.S. companies?
Ans: DeepSeek AI’s success has led to a significant drop in the stock prices of U.S. tech giants like Nvidia and raised questions about the costs involved in generating their AI modules. - Are DeepSeek’s AI models better than ChatGPT?
Ans: DeepSeek AI’s models perform like ChatGPT and are more affordable to use for certain tasks. - How does DeepSeek save money?
Ans: DeepSeek has used older computer parts and assembled the system in a way that makes everything work well and at a cheaper operating cost.