In a week where attention on DeepSeek’s low-cost AI models caused a bloodbath on Wall Street and propelled DeepSeek’s app to the top of iPhone download rankings in the U.S., China was celebrating. Emojis of “DeepSeek pride,” often with smiling cats or dogs, flooded Chinese social media, adding to the festive Lunar New Year atmosphere.
Chinese state media, as well as tech and business leaders, raved about DeepSeek.
“We need to believe that the moon in other countries is not necessarily rounder: Whatever others can do, we can also do it and even do it better,” read a profile of DeepSeek, published by the government of Zhejiang province where the company is located.
Local media reported that the small Guangdong province hometown of DeepSeek founder Liang Wenfeng has become a hot destination for tourists during the weeklong Lunar New Year holiday.
In China, the buzz around DeepSeek was a timely confidence boost as the Chinese leadership faces prolonged economic gloom and potentially another trade war with a Trump administration threatening more tariffs.
Feng Ji, co-founder of Game Science, the developer of popular videogame “Black Myth: Wukong,” hailed DeepSeek’s advancement as one that can change “a nation’s fortune.”
“So lucky, so happy! Such a shocking breakthrough came from a pure Chinese company,” he wrote on the Chinese social-media platform Weibo.
Several state-media outlets, including China Daily, went so far as to repost a clip from “The Daily Show” where host Jon Stewart facetiously asked: “Has Chinese AI put American AI out of a job?”
Zhou Hongyi, chairman of Chinese internet-security firm Qihoo 360, said China has already caught up with the U.S. in the AI race. “DeepSeek will definitely be a key member of China’s ‘Avengers team’” in countering the U.S.’s AI tech dominance,” he said in a video posted on his verified account on Weibo.
U.S. export controls on high-technology equipment and high-performance chips to China have hampered China’s ability to advance its technologies, although there have been loopholes that allowed Chinese companies to get some access to such technology and chips. Washington is trying to close those loopholes.
Beijing has sharply criticized the controls, calling them malicious attempts to suppress China.
Elsa Kania, an adjunct senior fellow at the bipartisan think tank Center for a New American Security, said Beijing’s rhetoric around DeepSeek’s success suggests a sense of pride and vindication, with state media conveying it as a triumph that a Chinese company has achieved technology advances despite U.S. export controls.
“But there’s also a sober recognition that U.S. export controls that have been trying to exploit chips as a chokepoint will continue to have major impacts,” Kania said. “This is a breakthrough, but not a game changer in terms of the fundamentals of the competition.”
ChatGPT maker OpenAI said it is investigating whether DeepSeek trained its new chatbot by repeatedly querying the U.S. company’s AI models. DeepSeek hasn’t commented.
Also discussed is DeepSeek’s reticence to dive into topics that are politically sensitive in China. When The Wall Street Journal this week asked it questions on topics that are routinely censored within China, including the so-called taboo “three Ts”: Tiananmen Square, Taiwan and Tibet, it returned more limited and often propaganda-infused answers.
Meanwhile, an investigation by Cyabra, a disinformation-detection firm, found that some of the hype about DeepSeek was driven by accounts it identified as fake. Cyabra, which examined conversations around DeepSeek on several social-media platforms during the week ended Tuesday, said it had identified 12% of the profiles engaging with the topic on X as fake, using tactics such as synchronized posting and repeating content.
Cyabra said the patterns and tactics employed by the fake profiles closely align with those observed in previous influence operations linked to China. Rafi Mendelsohn, Cyabra’s chief marketing officer, said there is a very strong chance that these accounts were created by the Chinese government to increase positive discourse around DeepSeek and drown out negative content.
A spokesman at the Chinese Embassy in Washington dismissed allegations of a state-led effort to promote positive DeepSeek content as groundless. “We always oppose attempts to manipulate social media to spread disinformation,” said the spokesman, Liu Pengyu.
After DeepSeek said on Monday that it was the victim of “large-scale malicious attacks,” Chinese social media was rife with speculation that the U.S., or DeepSeek’s U.S. rivals, were behind the attack. Some analysts have said the disruption to DeepSeek’s services could have been caused by a rise in user demands. DeepSeek said on its website Thursday that a fix had been implemented.
Write to Liyan Qi at Liyan.qi@wsj.com