DeepSeek's Revenue Soars: Annualized Run Rate Tops $10 Billion
DeepSeek has achieved a significant financial milestone, with its annualized revenue run rate now exceeding $10 billion. This represents a remarkable doubling of revenue in just a few months, up from less than $5 billion earlier this year. The surge is attributed to two primary factors: recent adjustments to API pricing and sustained strong market demand for its large language models.
Funding Round and IPO Plans Gain Momentum
During a recent investor meeting, CEO Liang Wenfeng shared these impressive financial results. He also revealed that the company is actively advancing its second major funding round while preparing for an initial public offering on the Shanghai Stock Exchange. The robust revenue growth is expected to bolster investor confidence as the company moves forward with these strategic initiatives.
DeepSeek aims to complete its current funding round by the end of October, targeting 50 billion yuan (approximately $7.5 billion) in new capital. If successful, the company would reach a valuation target of 500 billion yuan. These capital market activities demonstrate DeepSeek's preparation for intensified competition and expansion in the global AI landscape.
Price Adjustments Meet Sustained Market Demand
According to informed sources, part of DeepSeek's revenue growth stems from price adjustments implemented last month. The company increased model calling prices by factors ranging from 2.3 to 4.5 times previous rates. Notably, Liang told investors that these price hikes haven't led to customer attrition, with user demand remaining strong.
This response suggests growing market willingness to pay for high-quality AI models and indicates that DeepSeek's products have established significant value within customer workflows. The favorable price elasticity provides solid support for the company's future profitability prospects.
Strategic Resource Allocation: Heavy on R&D
Liang also detailed the company's resource allocation strategy to investors. DeepSeek continues to dedicate the majority of its resources to new model development, with over 70% of computing power allocated to model training. Less than 30% remains for inference operations of existing models.
This "R&D-first" resource distribution reflects the company's emphasis on technological innovation and long-term competitiveness. Liang noted that internal testing shows the company's lightweight models can run stably on gaming graphics cards while handling most everyday user tasks. This technical breakthrough not only reduces inference costs but also paves the way for broader deployment across various devices.
As AI competition intensifies globally, DeepSeek is positioning itself as an increasingly significant player through its three-pronged strategy of revenue growth, capital market activities, and technological innovation. The upcoming funding round and IPO process will serve as important indicators for observing the development trajectory of China's AI industry.