The rapid surge in demand for cloud-based AI systems is placing a significant strain on infrastructure. In response, the Chinese company DeepSeek has introduced a pricing strategy to help distribute usage more evenly. Starting Wednesday, access to the DeepSeek V3 model for developers will be available at half price between 00:30 and 08:30 Beijing time.
Additionally, the R1 model, which powers the widely used DeepSeek chatbot, will be 75% cheaper during these hours when accessed via API. DeepSeek’s services have gained considerable popularity in China, even among competing AI developers like Tencent Holdings and Perplexity AI. Government agencies are also actively using DeepSeek, as local regulations prohibit Chinese officials from processing data outside the country. Meanwhile, governments in Italy and South Korea have restricted their officials from using DeepSeek altogether.
Impact on Users and Future Updates
To manage high daytime demand, DeepSeek has warned users about potential access issues. By reducing costs during nighttime hours, the company aims to distribute server loads more efficiently. This change may be especially beneficial for international customers in different time zones, who can take advantage of lower prices during their regular working hours, adds NIXSolutions.
DeepSeek’s approach highlights the ongoing challenges in balancing AI system accessibility with infrastructure limitations. We’ll keep you updated on any further developments regarding their pricing and availability strategies.