A massive cyberattack targeted at China’s AI start-up DeepSeek in recent days originated in the US, according to China’s state broadcaster.
The cyberattack on DeepSeek started on January 3 and reached a peak on Monday and Tuesday with a massive brute-force attack from US IP addresses, said Yuyuan Tantian, a social media account affiliated with China’s state broadcaster CCTV, on Wednesday.
DeepSeek last week launched a free and open sourced AI assistant that claimed to use less data at a fraction of the cost of incumbent US artificial models, which was regarded by some as a “Sputnik moment” for America’s AI industry for possibly marking a turning point in the level of investment needed for AI.
Do you have questions about the biggest topics and trends from around the world? Get the answers with SCMP Knowledge, our new platform of curated content with explainers, FAQs, analyses and infographics brought to you by our award-winning team.
DeepSeek said on Monday it would temporarily limit new registrations to users with mainland Chinese mobile numbers because of a “large-scale malicious attack” which had resulted in problems in registration.
The earlier stage of the cyberattack contained more distributed denial-of-service (DDoS) attacks that aimed to disrupt DeepSeek’s normal service by overwhelming its servers and bandwidth with a flood of internet traffic, and the more recent attacks were primarily brute-force attacks, aiming to crack user ID and passwords in an effort to understand how DeepSeek works, said CCTV, quoting a report from China’s cybersecurity company QAX Technology Group.
A brute-force attack will systematically check all possible passwords and passphrases until the correct one is found. With the compromised ID and passwords, the attacker can pretend to be the registered users of web services to use and analyse their services.
“All the attack IPs were recorded, all are from the US,” Wang Hui, a QAX cybersecurity expert, told CCTV.
The report of US-originated cyberattacks on the Chinese platform comes days after the stunning launch of its DeepSeek-V3 open-source model, which its creators say is capable of rivalling the most advanced closed-source models globally with a fraction of the development cost.
The AI application has surged in popularity among US users since it was released on January 10.
DeepSeek’s new product created an earthquake on US tech stocks this week when American AI experts tried to decode DeepSeek’s formula for success.
Its impact also went beyond Wall Street and Silicon Valley, as the Trump administration vowed to pursue new restrictions and enforce them to keep the US’ lead in the global AI race.
Howard Lutnick, President Donald Trump’s nominee for Secretary of Commerce who would oversee future export controls on AI technology, told a Senate confirmation hearing on Wednesday that it appeared DeepSeek had misappropriated American AI technology and he vowed to impose restrictions.
“I do not believe that DeepSeek was done all above board. That’s nonsense,” Lutnick said. “I’m going to be rigorous in our pursuit of restrictions and enforcing those restrictions to keep us in the lead.”
David Sacks, the White House’s AI and crypto chief also raised concerns about DeepSeek distillation in a Fox News interview on Tuesday.
The distillation process uses an older, more established and powerful AI model to evaluate the quality of the answers coming out of a newer model, effectively transferring lessons from older models to the new one, allowing the newer model to reap the benefits of massive investment in the old models without incurring similar costs.
More from South China Morning Post:
- How tech start-up DeepSeek emerged as the unlikely game changer in US-China AI war
- Microsoft, OpenAI investigate China’s DeepSeek over data breach
- China’s DeepSeek is redefining AI tech. Here’s why the US must get on board
- White House evaluates effect of China AI app DeepSeek on national security
- DeepSeek’s rattling of US tech stocks could change how America sanctions China: analysts
For the latest news from the South China Morning Post download our mobile app. Copyright 2025.
