According to Dongcha Beating monitoring, the first round of 50 billion RMB financing for DeepSeek is nearing completion. In this round, Alibaba, Tencent, and the National Venture Capital Fund each contributed 10 billion, with founder Liang Wenfeng personally contributing 20 billion, bringing the company's valuation to approximately 3.5 trillion RMB. It is rumored that Huawei also participated in the investment negotiations, while traditional top VCs and PEs were unable to secure entry.
Initiating external financing has directly alleviated the dual talent and computing power crisis facing DeepSeek. Previously, core team member Guo Daya left due to salary inversion and insufficient computing power to join Byte. To stabilize the team, DeepSeek has doubled its current compensation, with core researchers' stock options reaching 8 digits. On the computing power side, although the V4 model's actual capabilities are approaching GPT-5.4, the underlying adaptation with chips like Huawei Ascend is still challenging, with some operators even needing DeepSeek staff to assist Huawei in coding. The company urgently needs more funds to directly purchase high-end computing power.
With the massive funding, DeepSeek is transitioning from a purely research institution to AI product development and commercial deployment. The company is accelerating its iteration pace, with plans to release the multi-modal large-scale V4.1 in June. Recent recruitment needs also show a clear preference for talent in multi-modal and Agent focusing on commercial deployment.
