On March 31, reporters, upon inspecting DeepSeek's official status page, discovered that from March 29 to 31, services such as web-based conversation, the mobile App, and API provided by DeepSeek encountered abnormalities for three consecutive days. The durations of these malfunctions were roughly 1 hour and 48 minutes, 10 hours and 13 minutes, and 1 hour and 3 minutes, respectively. Presently, all three issues have been rectified, and all services have returned to stable operation. According to official monitoring data, over the past 30 days, the overall availability rate of DeepSeek's web-based conversation service has stood at 98.61%. Lately, the successive service disruptions at DeepSeek have once more focused industry attention on the stability of AI infrastructure.
