PEIJING – Chinese companies use the new Deepseek artificial intelligence model to see how it can improve productivity.
The Chinese AI model has stormed the world in recent weeks after having presented its reasoning process and claims that the Openai chatgpt competes on the cost – despite the American restrictions on Chinese access to advanced semiconductors necessary to develop technology.
Eight car manufacturers, including Bydat least nine financial securities companies, Three telecommunications operators belonging to the State and smartphone brand Honor There are among the many who rushed last week to integrate into Deepseek. Cloud computing operators Alibaba, Huawei, Tencent and Baidu have all offered customers to access the latest Deepseek model.
“It’s completely unprecedented,” said Wei Sun on Monday, main artificial intelligence analyst of Counterpoint Research. It highlighted the adoption rate, the scale of enterprise integration and the extent of specific industries covered.
“When we have it all, we know that it has a big social and economic impact,” she said.
Optimism on artificial intelligence has spread to Chinese actions. UBS said on Wednesday that Chinese AI actions have increased by 15% since the start of the year, overputing the wider MSCI China index by 9%.
An important factor in widespread interest is timing. Deepseek published its latest R1 model on January 20, and the news of its low -cost reasoning capacities caused a world sale on technological actions on January 27 – just like Millions of urban workers in China returned to their hometown To celebrate the eight -day lunar New Year Holidays.
Consequently, less developed parties of China have acquired a better understanding of AI and its impact, a subject previously limited to conversations in the largest cities of China, said Wenhao Zhang, CEO of the marketing advice Consumption based in Beijing Doodod.
“It is a major market education.
Zhang, who studied AI at Tsinghua University, founded Doodod in 2012 to create customer engagement through social media analysis. He said that the company – which counts China Merchant’s Bank and Toyota as customers – began to watch the deepseek offers at the end of last year, and began to use it more after the R1 end January.
Deepseek, founded in 2023 of a quantitative specitative fund, had published a basic version of R1 in Novemberand a V3 model in December. He launched a Chatbot smartphone application in January.
Open source local deployment
Another attractive factor for companies is that Deepseek models are open-source, allowing individuals and companies to download and personalize it.
Deepseek also announced a considerably drop in applications prices to use its technology compared to that of Openai. Chatgpt is not officially available in continental China and obliges users to provide a telephone number and payment method abroad from a supported countries like the United States
Deepseek has changed the perception that AI models belong only to large companies and have high implementation costs, said James Tong, CEO of Movitech, a company software company that affirms that its customers Include Danone and the Chinese state network.
He said Movitech began to integrate a previous version of Deepseek in the fourth quarter of last year, helping to increase sales by about 25% compared to the same period in 2023. Customers’ ability to make decisions, he said.
Many recent videos on Chinese social media have shown how to manage a local version of Deepseek on Apple Mac Mini.
Apple Mac Mini online sales in China climbed considerably from November to January, compared to the same period of the previous year,, According to data from the WPIC council. The JD.com site focused on electronics recorded unit sales of approximately 20,200 in January, against almost 19,400 in December and around 12,250 in November, showed data.
Deepseek’s affordability is using more expensive AI models to reduce prices, allowing more companies to adopt technology, said Chim Lee, Senior Analyst in Asia economy. He added that open -source models allow financial, banking and health care companies – which are Subject to strict data protection rules in China – to develop AI applications locally.
“It is still very early to point out concrete commercial applications, but a key to remember is that Deepseek will accelerate the merchant of the AI,” said Lee.
Beijing also increases support. The national Chinese supercompute network announced on Tuesday that eligible companies and individuals Can obtain three months free to deeply access, as well as the subsidized calculation power.
The network is similar to Stargate project supported by Trump of Openai In the United States, for the construction of AI infrastructure – with the potential for “even faster scaling,” Winston Ma, auxiliary professor at the Nyu School of Law on Wednesday. He is also the author of “The Digital War: How China’s Tech Power shapes the future of AI, Blockchain and Cyberespace”.
Not centered on Deepseek
The precipitation of trying Deepseek does not mean that it will be the only AI supplier for Chinese companies. In the United States and China, developers regularly publish new models.
Movitech also uses the Qwen AI model of Alibaba, said Tong, noting that the market wants technology to reduce costs and produce the most results, whether Openai or Deepseek.
Hanghang AI, which has invested several hundred millions of yuan to develop AI solutions for companies in 20 industries, uses a range of models, said its partner and COO Shu Weibing.
Many people first used “Baidu, then realized that it was not as good as Kimi, so it was not as good as [ByteDance’s] Doubao, who also reduced prices, “said Shu in Mandarin, translated by CNBC.” Now it’s Deepseek. “”
It remains to be seen how much the generator can increase productivity and profits.
Shu provides that small businesses and companies that integrate AI into equipment will benefit more than large consumer-facing internet platforms, whose work of AI to date, he said, has focused more On the increase in efficiency rather than the creation of new consumption services.
Despite the drop in prices for AI models, “small and medium -sized enterprises can always be in a waiting period” to adopt technology due to the relatively high cost of a complete deployment, including power and personalization IT, Mike Fang, principal director analyst at Gartner, said Mandarin on Wednesday translated by CNBC.
But the consulting company provides that in 2027, the average price to access a generative AI model will be less than 1% of what it currently costs – and that in 2029, 60% of Chinese companies will have incorporated the ‘AI in their main products and services, forming the main drivers of income growth.
