DeepSeek Annualized Revenue Tops $1 Billion; Customers Stay After High-Margin Price Increase

On September 23, 2026, The Information reported that DeepSeek CEO Liang Wenfeng disclosed to investors that the company's annualized revenue run rate had surpassed $1 billion, doubling from several months earlier.

Factual Recap

According to The Information's exclusive report, Liang Wenfeng confirmed the revenue figures in communications with investors. On August 26, the outlet had reported that DeepSeek's revenue for the first seven months of this year was about RMB 475 million, roughly 10 times its full-year 2025 revenue. The September report further showed that over the past several months, the company's annualized revenue run rate doubled from its previous level to $1 billion.

The report also mentioned that DeepSeek adjusted its model prices in August, with increases ranging from 2.3x to 4.5x, but customer demand did not show a significant decline after the adjustment. The company's API prices remain in the lower range among major AI models.

Mechanism Breakdown

The annualized revenue run rate doubling from several months earlier to $1 billion shows that DeepSeek has made progress in both expanding revenue scale and controlling costs. Demand was unaffected after the price increase, indicating that its products' balance between performance and price has won user recognition. The report noted that the company's models can achieve performance comparable to U.S. rivals while using fewer Nvidia chips, providing a basis for maintaining relatively high gross margins.

Meanwhile, the report mentioned that Chinese AI labs, benefiting from higher model efficiency and lower domestic energy costs, can offer services at lower prices than U.S. companies. This is consistent with DeepSeek's current low-price positioning and also explains why customer retention remained stable after the price increase.

Industry Impact

DeepSeek is advancing a second funding round, aiming to raise RMB 5 billion at a valuation of RMB 50 billion, and plans an IPO on the Shanghai Stock Exchange. If the funding proceeds as planned, it will provide a new reference point for the performance of Chinese AI companies in the capital markets. The report showed that the company attracted attention from Silicon Valley and Wall Street early on by launching models with performance close to U.S. rivals but using fewer chips.

This funding round and IPO preparation are taking place against the backdrop of continuously rising AI costs. The report cited the Financial Times as saying that high AI costs have created opportunities for Chinese labs, enabling them to attract cost-conscious enterprises with lower prices.

Strategic Assessment

(The following is analysis, not fact.) Based on current revenue growth and stable demand after price increases, DeepSeek's path shows the possibility of balancing scale and profitability through efficiency advantages. This contrasts with the model of some U.S. AI companies that rely on large-scale investment for expansion, but the specific differences still require more public data to assess. If Chinese open-source models can continue to maintain similar gross margins and customer retention, this may affect the standards for evaluating cost and performance in global AI competition.

(The following is analysis, not fact.) However, the report did not provide specific revenue or cost figures for OpenAI or Anthropic, so the differences in their business paths cannot be directly quantified. Whether DeepSeek can turn its current revenue momentum into long-term sustainable profitability still depends on the execution of subsequent financing, the IPO process, and market competition dynamics.