The globally popular DeepSeek has been subjected to malicious cyber attacks.
According to reports, on January 28, the official website of DeepSeek indicated that its online services were under a massive malicious attack. Security experts from Qi An Xin revealed that the IP addresses involved in the cyber attack on DeepSeek were all located in the USA.
On January 29, reporters from The Paper learned from Qi An Xin that recent monitoring detected that DeepSeek has been under numerous overseas attacks for nearly a month, with tactics escalating since January 27. In addition to DDoS (Distributed Denial of Service) attacks, analysis also uncovered a large number of brute force attacks, putting DeepSeek's AI services and data through an unprecedented security test.
Experts from Qi An Xin XLab laboratory told The Paper that the attacks are expected to continue.
In response to the sudden escalation of massive DDoS attacks on the nights of January 27 and 28, DeepSeek responded and handled the situation promptly. At 00:58 AM on January 28, an IP switch was made during a powerful and destructive HTTP proxy attack initiated by the attackers.
Security experts from Qi An Xin told reporters that this large-scale attack is not an isolated incident. In recent years, cyber attacks against high-tech companies have been intensifying. The motives of the attackers are complex, ranging from commercial competition, attempts to steal core technological data, to even some state-backed hacker organizations trying to curb the development of China's high-tech industry through attacks.
According to the Global Times, the AI large model DeepSeek developed by a Chinese company has recently attracted global attention. On January 28, CNBC reported that the US Navy issued a warning letter, instructing its personnel not to use the DeepSeek model.
The warning letter from the US Navy was issued on January 24, four days after the launch of the reasoning large model DeepSeek-R1. A CNBC reporter who saw the email stated that the US Navy has instructed its troops and staff not to use the DeepSeek model in any way, including processing work or personal daily tasks, nor to upload, download, or transmit any information on the platform.
On New Year's Eve, DeepSeek quietly released the open-source multimodal model Janus-Pro. Previously, on the morning of January 27, the DeepSeek app topped the free app download charts in both the Apple China and US regions, surpassing ChatGPT in the US download rankings.
According to NBC, US President Trump stated at a Republican conference in Miami, Florida, on January 27 that the AI technology of the Chinese AI startup DeepSeek has sounded an "alarm" for US tech companies, emphasizing that "American companies need to focus on competition to win."
