DeepSeek is rapidly expanding its team and pitching investors a new growth plan, Ifeng Tech reports

DeepSeek is rapidly expanding its team, Chinese tech outlet Ifeng Tech reported on October 27, citing several people familiar with the matter. According to a person close to the company, DeepSeek is outlining a new growth roadmap to investors that includes a large expansion of its team. It comes as DeepSeek closes in on a funding round of about 100 billion yuan, which Reuters reported on October 22.

DeepSeek had already signaled its plans. On June 25, it said it would at least double the size of every department and advertised 33 positions spanning research, computing infrastructure and product development, including data-center construction roles in Ulanqab. In September, Cui Tianyi, head of the Harness team that turns DeepSeek’s models into agents, posted about 150 openings and called the hiring demand “unprecedented.” According to Yicai, Cui tied the drive to sharp growth in data volumes, training workloads, users and request traffic, warning that existing backend systems may not keep up.

Ifeng also says DeepSeek has announced it will open-source DeepInfer, its internal inference framework, compatible with both Nvidia and Huawei Ascend hardware. Per the report, OpenRouter data shows DeepSeek’s models topping the platform’s weekly usage rankings for three straight weeks. Founder Liang Wenfeng said in early 2025 that the team had just 160 people, and no newer headcount has been disclosed. The reports include no comment from DeepSeek.