At the beginning of 2025, the Chinese AI app DeepSeek-R1 suddenly became the most popular free app in the US, causing a major stir in Silicon Valley and the stock market. Claims that the app competed with ChatGPT and was developed at low cost led to Nvidia’s market value dropping by $600 billion.
DeepSeek questioned the need to rely on large and expensive infrastructure by offering small, efficient models. Some American startups continue to use the app to reduce costs but do so cautiously due to security concerns.
In the AI rivalry between China and the US, DeepSeek disproved the view that China was behind and showed new possibilities for creating cheap, competitive models. At the same time, the US government is concerned about DeepSeek’s ties to Chinese military and intelligence agencies.
Currently, OpenAI and other major US companies are investing more in large models and powerful computing infrastructure. The release of GPT-5 and the recovery of Nvidia’s market value show that resource-intensive approaches still dominate AI.
DeepSeek faces challenges releasing its second-generation DeepSeek-R2 model due to chip shortages and intense competition. Ultimately, DeepSeek’s initial revolutionary impact was temporary, and AI development continues along a resource-heavy path.
Read 'Zamin' on Telegram!
