According to two people familiar with the matter, DeepSeek's current annualized revenue operating rate has reached 1 billion US dollars. Just a few months ago, the company's target was less than $500 million. On the one hand, the increase in revenue was due to the recent increase in API pricing, and on the other hand, it also stemmed from the continued popularity of its large model in the market. DeepSeek CEO Liang Wenfeng revealed the latest revenue data at a recent investor conference. The company is advancing the second round of financing and is preparing to be listed on the Shanghai Stock Exchange. This revenue performance is expected to boost investor confidence. DeepSeek plans to complete this round of financing by the end of October, with a target of 50 billion yuan and a valuation target of 500 billion yuan. According to people familiar with the matter, part of DeepSeek's revenue growth came from last month's price adjustment action: the model call price was raised to 2.3 times to 4.5 times the original price. Liang Wenfeng said at the meeting that price adjustments have not caused loss of customers, and user demand is still strong. People familiar with the matter quoted Liang Wenfeng as revealing to investors that DeepSeek continues to invest most of its resources in developing new models. The company allocates more than 70% of the computing power to model training, leaving less than 30% of the computational power for inference operations of existing models. Liang Wenfeng also told investors that internal tests showed that the company's lightweight model can run stably on gaming graphics cards and can handle the vast majority of users' daily tasks.

According to two people familiar with the matter, DeepSeek's current annualized revenue operating rate has reached 1 billion US dollars. Just a few months ago, the company's target was less than $500 million. On the one hand, the increase in revenue was due to the recent increase in API pricing, and on the other hand, it also stemmed from the continued popularity of its large model in the market. DeepSeek CEO Liang Wenfeng revealed the latest revenue data at a recent investor conference. The company is advancing the second round of financing and is preparing to be listed on the Shanghai Stock Exchange. This revenue performance is expected to boost investor confidence. DeepSeek plans to complete this round of financing by the end of October, with a target of 50 billion yuan and a valuation target of 500 billion yuan. According to people familiar with the matter, part of DeepSeek's revenue growth came from last month's price adjustment action: the model call price was raised to 2.3 times to 4.5 times the original price. Liang Wenfeng said at the meeting that price adjustments have not caused loss of customers, and user demand is still strong. People familiar with the matter quoted Liang Wenfeng as revealing to investors that DeepSeek continues to invest most of its resources in developing new models. The company allocates more than 70% of the computing power to model training, leaving less than 30% of the computational power for inference operations of existing models. Liang Wenfeng also told investors that internal tests showed that the company's lightweight model can run stably on gaming graphics cards and can handle the vast majority of users' daily tasks.