Sources: DeepSeek's annualized revenue recently reached $400M to $500M as it seeks to raise ~$7.4B at a ~$74B valuation, up from ~$50B after raising ~$7B in May
The Information:
Context & Ripple Effects
DeepSeek’s financing story has moved rapidly from early discussions about an outside round and a $10B-plus valuation in April to a roughly $7B first raise at more than $50B by late May. Related coverage now places the company in talks for still more capital at about a $71B valuation.
The reported $400M–$500M annualized revenue gives investors a revenue marker alongside that valuation escalation. The company is also described as expanding infrastructure, considering an IPO in China, and developing inference hardware intended to lessen dependence on Nvidia and Huawei.
First-order effects
- DeepSeek can use a new roughly $7.4B round, if completed, to extend its infrastructure buildout and fund its stated push into inference-chip development.
- The higher proposed valuation and reported revenue run rate give DeepSeek a stronger basis for investor discussions and for IPO planning, while raising expectations for its execution after the May financing.
Second-order effects
- DeepSeek’s infrastructure spending would increase pressure on its existing hardware suppliers, while its chip effort creates a potential future alternative to continued purchases from Nvidia and Huawei.
- AI rivals competing for researchers and capital face a more highly valued DeepSeek with greater capacity to retain talent and finance compute-intensive product development.
Third-order effects
- If large private rounds continue to accompany rising revenue at AI model developers, the sector may become more concentrated around companies able to finance both model development and the underlying inference stack.
- DeepSeek’s reported IPO preparation and hardware ambitions point toward a model-provider strategy that spans financing, infrastructure, and silicon; whether that integration reduces supplier dependence will depend on execution rather than fundraising alone.
The trend: This is part of the shift from AI-model companies raising to build models toward heavily financed, vertically integrated operators investing in revenue scale, inference infrastructure, and hardware control.