DeepSeek Hits $1B Run Rate Ahead of $7.45B Funding Round

DeepSeek annualized revenue has topped $1 billion, driven by model pricing increases of 2.3 to 4.5 times, as Chief Executive Liang Wenfeng disclosed the milestone during a recent investor meeting ahead of a planned $7.45 billion fundraising push.

### Revenue Scale and Pricing Power

The Chinese AI developer is preparing for a second funding round targeting 50 billion yuan, equivalent to about $7.45 billion, aiming for a 500 billion yuan valuation by the end of October. At the same time, the company is laying groundwork for a potential listing on Shanghai's STAR Market. While the $1 billion run rate marks significant commercial traction, it highlights a continuing revenue gap relative to U.S. giants such as Microsoft-backed OpenAI and Anthropic.

### Compute Allocation and Model V4.1-Flash

Operational capacity remains heavily tilted toward research and development. More than 70% of DeepSeek computing capacity is dedicated to training new models, with under 30% supporting existing-model inference. Earlier in September 2026, the company introduced V4.1-Flash, engineered for reduced memory and storage overhead to improve serving economics during high-throughput tasks.

### Money Play

As international AI infrastructure competition evolves, monitor pricing sustainability and compute allocation shifts between training and commercial inference workloads.