DeepSeek, the rapidly growing Chinese AI startup, has reopened access to its core programming interface (API) after a nearly three-week hiatus. The suspension, which began in early February due to capacity shortages, had temporarily halted customers from topping up credits for use with the platform. With the service now back online, the 20-month-old company is looking to support growing demand for its powerful AI tools, which have gained significant traction since their debut last month.
A representative from DeepSeek confirmed in a verified company chat on WeChat that while top-ups have resumed, server resources will remain limited during peak daytime hours. This announcement comes as the company works to scale its infrastructure to meet the surging interest from both domestic and international markets.
The reopening of access coincides with heightened competition in China’s rapidly evolving AI industry. On the same day, Alibaba Group unveiled a preview of its latest AI model, QwQ-Max, signaling the tech giant’s continued push to challenge DeepSeek’s rising dominance. Meanwhile, ByteDance, the parent company of TikTok, revealed it is testing its own deep reasoning AI model with a select group of users, adding another layer of competition for DeepSeek.
Alibaba, in a bold move to strengthen its AI capabilities, announced plans to invest $53 billion over the next three years to enhance its cloud computing and AI infrastructure. The company also revealed intentions to open-source QwQ-Max, a strategic step aimed at broadening its AI ecosystem and attracting developers globally.
Since its launch, DeepSeek has become a key player in the Chinese tech scene, sparking a stock market rally across mainland China and Hong Kong. The startup’s artificial intelligence chatbot, developed at a fraction of the cost of its competitors, has been widely adopted by Chinese businesses across various sectors. Despite its growing popularity, DeepSeek has faced scrutiny from foreign governments, with some countries—ranging from Australia to the US—moving to block its usage due to national security concerns.
In an unexpected move, DeepSeek also revealed plans last week to release key parts of its code and data to the public, setting the stage for a more open approach to AI development. This decision puts the startup on a collision course with rivals like OpenAI, which have been more cautious about sharing their core technology.
Taking its first steps toward this transparency, DeepSeek announced on Wednesday that it would make its “DeepGEMM” library available. This resource, used in the training and inference of its V3 and R1 AI models, represents the beginning of the company’s broader effort to share its technological advancements, further intensifying the race between the US and China to dominate the next generation of AI.
Related topic:
China’s Property Sector Faces Gloomy Outlook Amid Slower Recovery
Markets Struggle Amid U.S.-China Tensions and Weak Economic Data
