The artificial intelligence open-source model developed by Chinese startup DeepSeek has made AI services more accessible and affordable, reshaping the foundations of many industries, said industry insiders.
Released on Jan 20, it has quickly soared to the top of Apple's app store's free charts by Jan 27, surpassing OpenAI's ChatGPT.
According to DeepSeek, in tasks such as mathematics, coding and natural language reasoning, the performance of this model is comparable to the leading models from heavyweights like OpenAI but only at a fraction of the cash and computing power of its competitors.
According to incomplete statistics, at least 18 domestic and foreign cloud service-related companies have deployed DeepSeek.
The Ministry of Industry and Information Technology (MIIT) said on Feb 8 that China's three telecom giants -- China Telecom, China Mobile and China Unicom have fully accessed the DeepSeek open source model.
"The open-source nature (of DeepSeek) brings a sea change to the entire AI industry. Our services have covered more than 300 (computing) nodes across major cities nationwide," said Feng Qiang, deputy general manager of the cloud computing division at China Unicom Digital Technology.
"Our intelligent computing platform has connected to 50 computing power partners at the same time, and has launched support for DeepSeek-R1 and V3 (versions) in multiple (computing) nodes across the country," said Lin Binghuai, chief research and development expert of China Telecom's Tianyi Cloud.
Industry insiders said that DeepSeek makes it possible for users to access more AI services at lower costs.
"The cost of the DeepSeek big model's training and reasoning has been greatly reduced. It only costs 0.8 yuan (about 0.11 U.S. dollars) to input 1.5 million words. On the first day of DeepSeek's launch, 15,000 new users came to use it, which is a staggering number," said Xin Zhou, general manager of Chinese tech giant Baidu's cloud platform Baidu Intelligent Cloud Qianfan Large Model Platform.
In addition to cloud service providers, hardware manufacturers have also embraced DeepSeek.
Data show that more than 15 chipmakers have adapted their products to the DeepSeek model service, and DeepSeek-powered intelligent computing machines are being launched.
Ouyang Jian, CEO of Beijing-based chipmaker Kunlunxin, said it has completed the adaptation to DeepSeek by themselves in less than a week, while previously, it took a long time to adapt to other large language models which are not open source, as they needed to communicate and collaborate with the models' developers.
"For open source models, they naturally provide us with many algorithm details and calculation process details. It not only takes less time (to adapt to the models), but also makes it easier to achieve a higher computing efficiency," said Ouyang.
In addition, demand for DeepSeek-integrated intelligent computing machines has also increased significantly across various industries.
"It is similar to buying a computer. People can use the machine right after unpacking it. State-owned enterprises and universities have a strong demand for such DeepSeek-integrated intelligent computing machines," said Lin Binghuai, an expert with China Telecom.
DeepSeek makes AI more accessible, affordable
