DeepSeek's affordable design improves expect China AI revolution

DeepSeek stirs nationalistic fever amid Sino-U.S. competition

AI-related stocks in China and Hong Kong surge

By Samuel Shen and Jiaxing Li

SHANGHAI/HONGKONG, Feb 6 (Reuters) - Chinese financiers are hurrying into AI-related stocks, betting the expert system advance of home-grown startup DeepSeek will cause a boom in the sector and provide the initiative to China in an intensifying Sino-U.S. technology war.

Feverish purchasing has pumped up shares of Chinese chipmakers, software application designers and data centre operators amid patriotic calls for an upward repricing of Chinese assets as U.S. President Donald Trump recharges a trade war with fresh tariffs.

"DeepSeek's advancement shows Chinese engineers are creative and capable of creations that can take on Silicon Valley," said China Europe Capital Chairman Abraham Zhang. "It has also stirred nationalistic fever in capital markets."

DeepSeek shocked Silicon Valley and rocked Wall Street late last month with the announcement of a competitive big language model that was ostensibly less expensive to develop than those of big-spending U.S. leaders such as OpenAI and Meta.

The occasion was explained as a watershed moment by Huaxi Securities analysts and has since seen cash gushing into AI-related stocks in mainland China and Hong Kong.

The Hang Seng AI Index has jumped more than 5% today while indices tracking chipmakers and IT more than 11%, assisting constant the Hong Kong market as the U.S. included a 10% tariff to Chinese imports.

On the mainland, investors returning from a week-long Lunar New Year holiday on Wednesday likewise piled into the tech sector, improving shares of firms in AI, semiconductors, huge information and robotics.

"2025 will witness a surge of AI applications," said Zhou Yingbo, wolvesbaneuo.com head of investment at Futures Vessel Capital.

"We're very positive about opportunities created by this revolution," Zhou said, expecting extensive adoption of both AI hardware and software by customers and organizations alike.

Likely beneficiaries consist of Nancal Technology, Suzhou MedicalSystem Technology, Doctorglasses Chain, Bestechnic Shanghai and Ucap Cloud Details Technology, Huaxi Securities said.

The DeepSeek advancement highlights how the U.S. attempt to slow China's technological improvement "has backfired, instead accelerating Chinese AI development," TF Securities said in a customer note. It called for a repricing of Chinese technology stocks which have actually underperformed U.S. peers in current years amid increased regulatory examination and geopolitical tension.

The emergence of DeepSeek could trigger even tighter U.S. innovation export constraints however that will only invite more federal government assistance and turbo-charge development, the brokerage said.

Goldman Sachs expects Chinese developments in AI development and application "could materially modify" the stock market trajectory.

The Wall Street bank estimates AI-enabled performance improvement might increase profits by 2% for Chinese equities, while brighter growth potential customers could lead to a 20% appraisal uplift for Chinese firms, narrowing the space with U.S. peers.

China's "hard tech" stocks trade at a cost representing 23.6 times incomes, while "soft tech" shares trade at 13.9. The price-to-earnings ratio of the greatest U.S. tech stocks, the so-called "Mag 7", is 31, revealed the Goldman report dated Feb 4.

DeepSeek has produced such a buzz that Chinese business up and down the AI worth chain, from chipmakers to cloud service companies are checking out possibilities with the startup's low-priced services, including heavyweights such as Huawei Technologies, Alibaba and Baidu.

Yi Xiangjun, partner of Shenzhen Black Stone Asset Management, said he is "all in" China's AI and tech stocks, wagering big, successful business will emerge in what he called an epoch-making transformation.

However, Wang Zhuo, partner of Shanghai Zhuozhu Investment Management, was more cautious.

"Many companies are still far way from producing make money from AI ... As a worth financier, I do not feel great putting cash into these stocks." (Reporting by Samuel Shen and Jiaxing Li; Editing by Vidya Ranganathan and Christopher Cushing)