DeepSeek, the Chinese AI startup making news today with its effect on global stock markets, has temporarily restricted new user registrations after reporting “large-scale malicious attacks” on its services.
This development follows the release of its R1 reasoning model, which has disrupted the AI industry with its high performance and cost efficiency.
Trained on Nvidia H800 GPUs under U.S. sanctions, R1 matches or exceeds the benchmarks of models like OpenAI’s o1 while costing a fraction to develop. The company’s rise has drawn global attention, with its app now surpassing ChatGPT as the top download on Apple’s U.S. App Store.
The cyberattack compounds the pressures already mounting on U.S. tech giants like Meta, whose AI teams are reportedly in “panic mode” over DeepSeek’s achievements, as revealed in internal discussions shared online.
Meanwhile, DeepSeek’s success also underscores the limits of U.S. export controls, with founder Liang Wenfeng emphasizing the importance of efficiency under constraints.