Chinese AI startup DeepSeek has surpassed ChatGPT in downloads on the US App Store, triggering a decline in tech stocks while raising concerns about data privacy due to its data storage on Chinese servers.
DeepSeek, a rising player in artificial intelligence, has become the most downloaded free app in the US App Store since launching its AI Assistant on January 10, 2025. Founded in Hangzhou in 2023, the company has quickly gained prominence in the industry with its AI model, DeepSeek R1. The app’s rapid success has reshaped the competitive landscape of AI, challenging the dominance of US-based companies and casting doubt on the effectiveness of Washington’s export restrictions.
DeepSeek R1 was developed in just two months for a remarkably low budget of $6 million, significantly undercutting the billions invested by Western tech giants. Its operational costs are reportedly 98% lower than ChatGPT’s, making it an appealing alternative for users seeking efficiency at minimal expense. Despite US trade limitations on advanced semiconductor exports, DeepSeek utilized readily available Nvidia H800 chips alongside its proprietary innovations to power its AI model.
The app’s emergence as a serious contender in the AI industry has sent shockwaves through Silicon Valley. Western AI-related technology stocks experienced a sharp decline following DeepSeek’s surge in popularity, with investors reacting to the startup’s promise of high-performance AI at reduced costs. The market downturn has resulted in substantial financial losses for some of the world’s wealthiest individuals, amounting to $108 billion in total. The most significant impact was observed in Nvidia’s stock, which fell by 17%, wiping out a record $589 billion in market value—the largest drop in the company’s history.
As DeepSeek continues to gain traction, cybersecurity experts have raised concerns about the app’s data storage policies. Philippe Mazan, senior manager of advanced threat detection and AI research at ESET, has warned about potential privacy risks, emphasizing that DeepSeek stores user data—including text, audio, uploaded files, and chat history—on servers located in China. This raises critical questions about how such information might be accessed and utilized.
Mazan advises users to avoid sharing sensitive personal or professional data when using DeepSeek, as even seemingly harmless queries related to health or politics could pose security risks. He notes that privacy concerns apply to AI models across the board, including OpenAI’s ChatGPT, but the location of DeepSeek’s servers adds another layer of uncertainty regarding data usage and potential government access.
In addition to privacy risks, cybersecurity analysts caution that DeepSeek’s accessibility may expose it to misuse. The model’s capabilities could be exploited by cybercriminals to refine malicious software, generate misleading content, or manipulate public discourse. In the wrong hands, AI-powered tools like DeepSeek could contribute to the spread of disinformation or serve as instruments of digital warfare.
Despite these concerns, DeepSeek’s rapid ascent underscores the shifting dynamics of the AI industry. The startup’s success challenges the long-held perception of US superiority in artificial intelligence and signals the growing influence of Chinese innovation in the global tech sector.
