The investor group includes several major Chinese companies across manufacturing, technology and financial services. Midea, Meituan, Geely Automotive and semiconductor manufacturer GigaDevice are participating as new investors, while Boyu Capital, CPE and Legend Capital are also involved. Existing shareholder IDG Capital is participating alongside state-backed investment entities.
The financing has not been publicly confirmed by DeepSeek, and the final amount could change before the transaction is completed. DeepSeek, CATL and Tencent have not provided immediate responses to requests for comment.
The substantial fundraising comes as DeepSeek prepares for a possible public listing in China. In September, the company appointed CITIC Securities to assist with preparations for an initial public offering on the STAR Market, a Shanghai-based exchange focused on technology companies. No definitive IPO timetable or transaction structure has been announced.
DeepSeek has built its reputation around developing AI models that deliver competitive capabilities at substantially lower costs than those offered by leading American developers. Its R1 reasoning model attracted widespread attention in January 2025, prompting investors to reconsider assumptions about the computing resources required to develop advanced AI systems.
The company's latest fundraising underscores the capital requirements associated with expanding those capabilities. While DeepSeek has emphasized efficiency in model development and deployment, operating increasingly capable AI systems still requires significant investment in computing infrastructure.
DeepSeek is reportedly developing a data center equipped with at least 160,000 Huawei AI chips while also working on its own inference processor. These efforts come amid US export restrictions limiting Chinese access to advanced Nvidia semiconductors.
The company has also continued releasing models and infrastructure software designed to reduce operating costs. In September, DeepSeek introduced V4.1-Flash, an updated model focused on faster inference, greater processing capacity and improved performance at scale.
DeepSeek has additionally made its DeepInfer inference software publicly available, allowing other organizations to use technology developed for running its models more efficiently. Its V4-Pro preview has also been reported to cost approximately 98% less than OpenAI's GPT-5.5 Pro, while offering downloadable model weights under an MIT license.
Those releases reflect DeepSeek's effort to maintain its low-cost approach while expanding its technical capabilities. The new financing would provide additional resources for that development, although the company has not disclosed how the proceeds will be allocated.
The funding also highlights growing financial support for Chinese AI developers. DeepSeek's investor base now spans battery manufacturing, internet services, automotive production, semiconductors and investment management, bringing together companies with different interests in the development and deployment of artificial intelligence.
Despite the size of the proposed transaction, DeepSeek would still have raised substantially less capital than OpenAI and Anthropic, which have each reportedly secured more than $100 billion. The comparison illustrates the financial scale of the competition among major AI model developers.
For DeepSeek, the immediate priorities include completing the financing, continuing model development and advancing preparations for a potential IPO. The company has not confirmed when it might list or how its latest funding would affect those plans.
If completed at the reported amount, the transaction would rank among the largest fundraising rounds for a Chinese company, giving DeepSeek significantly more capital to expand its AI operations while continuing to compete on model performance and cost.
This analysis is based on reporting from AInvest.
Image courtesy of DeepSeek.
This article was generated with AI assistance and reviewed for accuracy and quality.
