- The latest funding round values the Chinese AI startup at about $75 billion, more than doubling its valuation in three months
- Revenue has surged despite a recent increase in API prices of up to 4.5 times
Chinese AI upstart DeepSeek has raised about $7.5 billion in a new funding round that values the company at roughly $75 billion, while its annualized revenue run rate has surpassed $1 billion, according to people familiar with the matter.
DeepSeek CEO Liang Wenfeng (梁文锋) disclosed the latest revenue figure at a recent investor meeting, saying the company’s annualized revenue had more than doubled from less than $500 million a few months ago.
The Information first reported the figure, citing two people with direct knowledge of the matter. Reuters said it could not immediately verify the report.
The revenue growth comes partly after DeepSeek raised prices for access to its AI models last month. Model API prices increased to between 2.3 and 4.5 times their previous levels, according to people familiar with the matter.
Liang told investors that the increases had not led to customer losses, with demand remaining strong.
Valuation jumps
The funding is DeepSeek’s second external financing round. The company’s first round, completed in June, raised about 50 billion yuan ($7.45 billion) and valued DeepSeek at more than 3.5 trillion yuan, according to Reuters.
The latest round had been reported to target about 50 billion yuan at a pre-money valuation of 500 billion yuan, with investors including semiconductor-focused Stone Creek Capital, SMIC-affiliated China Fortune-Tech Capital, Boyu Capital, CPE and a Hefei government investment platform.
Stony Creek Capital was reported to have put in about 5 billion yuan, Chinese business media outlet ChinaVenture reported on September 24.
The venture firm was jointly established by memory chip giant GigaDevice (兆易创新) and an investment team as a semiconductor-focused investment firm.
New heavyweight backer
GigaDevice founder Zhu Yiming (朱一明) is also the founder of CXMT (长鑫存储), one of China’s largest DRAM makers.
DeepSeek’s first round included a roughly 20 billion yuan commitment from Liang, 10 billion yuan from Tencent and 5 billion yuan from CATL, according to Reuters. The round also roped in investors including NetEase, JD.com and IDG Capital.
The latest valuation marks a sharp increase in just three months, underscoring the premium investors are placing on DeepSeek as it turns its open-source AI models into a growing commercial business.
DeepSeek is also preparing for a potential listing on the Shanghai Stock Exchange’s STAR Market.
The company has hired investment banks to work on the process and could file as early as the end of this year, according to Chinese media reports.
Training takes priority
Liang told investors that DeepSeek is continuing to devote most of its computing resources to developing new models.
More than 70% of the company’s computing capacity is allocated to model training, leaving less than 30% for inference, or running existing models for users.
Internal tests also showed that DeepSeek’s smaller models can run reliably on consumer gaming GPUs and handle most everyday tasks, Liang said.
Wider deployment across less specialized hardware could eventually reduce pressure on dedicated AI computing infrastructure.
The figures point to a business model that remains heavily centered on API sales rather than consumer subscriptions or advertising.
DeepSeek generated about 475 million yuan in actual revenue during the first seven months of 2026, while its annualized run rate has since climbed above $1 billion, The Information reported.
Header image credit: DeepSeek
