China’s AI start-up DeepSeek, a leading platform in advanced data analytics, has temporarily halted new user registrations following a “large-scale malicious attack” on its systems.
Founded in 2023, DeepSeek is committed to advancing artificial general intelligence (AGI). The company gained global recognition this week when its iOS app, DeepSeek—AI Assistant, became the top-rated free app on Apple’s App Store in the U.S., briefly surpassing OpenAI’s ChatGPT in popularity.
Late Monday, DeepSeek described the cyberattack as a “deliberate and sophisticated breach” aimed at compromising its infrastructure.
While the company confirmed that no customer data was accessed or compromised, the attack disrupted its registration systems, leading to a temporary suspension of new sign-ups.
“Due to large-scale malicious attacks on DeepSeek’s services, we are temporarily limiting registrations to ensure continued service. Existing users can log in as usual. Thanks for your understanding and support,” the company stated on its status page.
Users attempting to register encounter an error message stating that new sign-ups are temporarily unavailable due to the attack, with instructions to try again later. Some users also report seeing a message indicating that the registration system is experiencing high traffic.
DeepSeek’s security team is currently assessing the impact of the attack while working with external cybersecurity experts to investigate the breach and enhance its protective measures. Updates will be provided as the investigation progresses.
“We are prioritizing security enhancements and will fully restore functionality once we are confident in our defenses,” a company spokesperson stated.
Despite the incident, existing users remain unaffected and can continue using the platform without disruption. However, DeepSeek has urged its users to stay cautious, update passwords, and enable multi-factor authentication.
New users will have to wait until the company reinforces its security systems to prevent future threats.
Last week, DeepSeek launched its free, open-source reasoning model, DeepSeek-R1, as a competitor to OpenAI’s o1. The model quickly gained global attention for its cost efficiency and computing power.
The company’s flagship DeepSeek-V3 model, featuring over 600 billion parameters, powers its AI assistant. DeepSeek claims this open-source model can rival the world’s most advanced closed-source models. Notably, it was trained using Nvidia H800 chips, with costs kept under $6 million.
