DeepSeek eyes $15 billion funding round as valuation hits $74 billion

A massive capital infusion

DeepSeek is currently navigating a significant funding round that reflects its growing status as a major player in the artificial intelligence space. According to reports from Reuters and Bloomberg, the company is seeking roughly 100 billion yuan, which converts to about $15 billion, from a mix of new and existing investors. This amount was reached after the target was raised twice during the process, starting at 50 billion yuan and moving to 80 billion yuan before hitting the current mark.

The demand for equity in DeepSeek appears to have outpaced the company's initial expectations. This surge in interest has pushed the pre-money valuation of the firm to approximately $74 billion, or about 500 billion yuan. While the round has not formally closed as of late October, the final list of participants and specific terms remain subject to change.

The investor list includes some of China's most prominent industrial and technology giants. Major backers like Tencent and CATL are returning for more capital, joined by over 15 new participants including Geely, Midea, Meituan, GigaDevice, and NetEase. Other notable financial institutions involved include Boyu Capital, CPE, Legend Capital, and IDG Capital.

Corporate structure and control

DeepSeek was founded in 2023 by Liang Wenfeng, who originally funded the company through his quantitative hedge fund, High-Flyer. The company first took outside capital in June, raising about $7.4 billion at a post-money valuation of $50 billion. That initial round included participation from Tencent, CATL, JD.com, NetEase, and China’s national AI industry investment fund.

The structure of these investments is designed to protect the founder's control over the company. Capital flows into a limited partnership that Liang personally controls, providing external investors with economic upside and a five-year lockup period. However, these investors do not receive voting rights, allowing Liang to retain roughly 84% of the economic interest and near-total voting control.

Analysts suggest a similar structure may apply to the current $15 billion round, though these details have not been finalized. By serving as the general partner of the limited partnership, Liang maintains his grip on the company's direction. This setup allows DeepSeek to absorb massive amounts of capital from diverse industries without diluting the founder's ultimate authority.

Technical focus and revenue

Despite its rapid growth, DeepSeek remains heavily focused on the research and development side of artificial intelligence. CEO Liang told investors that the company currently devotes more than 70% of its compute resources to training new models. This means that less than 30% of its processing power is actually allocated to serving existing models to users.

The company recently demonstrated its technical progress by releasing DeepSeek-V4.1-Flash. This new model is designed to provide faster inference speed and higher throughput, which are critical metrics for how quickly an AI can generate responses. These technical milestones are being achieved while the company simultaneously builds a sustainable business model.

The financial health of the company appears to be strengthening alongside its technical capabilities. Reports indicate that DeepSeek’s annualized revenue run rate surpassed $1 billion in September. If the current funding round successfully closes, the company will have raised more than $22 billion in outside capital in the year 2026 alone.

Geopolitical and market friction

The rise of DeepSeek is occurring against a backdrop of significant geopolitical tension. Just days before the news of the funding round broke, the U.S. Department of Homeland Security ordered federal agencies to strip DeepSeek’s technology from their systems. This move highlights the growing divide between Western security interests and Chinese AI development.

While DeepSeek seeks to become a national champion in China, it faces a complex path toward public markets. The company has engaged CITIC Securities to organize a debut on the STAR Market in Shanghai, with a potential timeline starting in 2027. This move puts it on a similar trajectory to other Chinese AI labs that are racing to prove their commercial viability.

The path to a successful listing is not without risks, as seen by recent market performance in the region. For example, the share price of Zhipu AI increased by 524% over a 43-day period following its January debut, while MiniMax saw its valuation drop significantly after announcing its own plans for a mainland listing. These divergent paths serve as a cautionary sign for DeepSeek as it prepares for its eventual debut on the STAR Market.