[REVISION HISTORY]

DeepSeek valuation, fundraising and IPO preparations

20 recorded revisions of this situation.

What changed

2026-10-24 11:14 UTC → 2026-10-27 07:25 UTC · added ~~removed~~

By ~~August~~ October 2026, DeepSeek had ~~reached a valuation of~~ secured approximately $60 billion. CEO Liang Wenfeng attributed this success to a non-traditional management philosophy that rejects the intensive ‘996’ work culture, allowing researchers to spend half their workday on autonomous research. In late August 2026, the company sought to raise 50 $15 billion ~~yuan (≈$7.4 billion)~~ in a ~~new~~ massive second funding round, bringing its total external investment to roughly $22 billion. This round, which ~~could value the firm at approximately 500 billion yuan (≈$74 billion). Despite a net loss of 715 million yuan between January~~ involved over 15 investors, was led by major backers including CATL and ~~July 2026, revenue surged nearly tenfold to reach 475 million yuan during that period, maintaining a gross margin of 44.6%~~ Tencent, with additional participation from Geely, Midea, Meituan, and ~~API service margins~~ IDG Capital. The funding target reportedly grew from initial goals of ~~82.9%.~~ 50 billion and 80 billion yuan to approximately 100 billion yuan. As ~~IPO preparations~~ the company prepares for a potential IPO on the Shanghai STAR ~~Market intensified, DeepSeek planned to appoint Yan Wentao,~~ Market—with a ~~partner at GL Ventures,~~ target debut as early as 2027—it is aggressively scaling its ~~first Chief Financial Officer to transition from a research laboratory~~ operations. DeepSeek plans to ~~a conventional corporate structure. The~~ at least double the size of all departments, focusing on technical R&D, computing infrastructure, and frontier algorithm research. To support this growth, the company ~~also engaged CITIC Securities~~ is planning to ~~assist with~~ integrate Huawei Ascend 950DT chips into its data centers. On the ~~listing process. By late September 2026,~~ technical front, DeepSeek ~~reported~~ recently open-sourced its internal inference framework, DeepInfer, which is compatible with both NVIDIA and Huawei Ascend computing platforms. The company’s commercial momentum remains strong, with an annualized revenue run rate of $1 billion. This growth followed a strategic pricing adjustment where model costs were increased by 2.3 to 4.5 times, though management noted demand remained steady. The company continues to allocate over 70 percent of its computing power to training new models. By October 2026, investor interest culminated in a massive funding round totaling approximately 100 billion ~~yuan (≈$15 billion), doubling the initial fundraising target. Major participants include Tencent Holdings, CATL, Geely, IDG Capital,~~ and ~~others. This capital is intended to fuel competition with global leaders like OpenAI and Anthropic by supporting the development of GW-scale computing parks, proprietary data centers, and~~ reports from the ~~continued iteration of~~ OpenRouter platform indicating that DeepSeek models ~~such as DeepSeek-V4.1-Flash. This latest round brings~~ held the ~~company’s total external investment to roughly $22 billion.~~ top position for weekly model usage for three consecutive weeks.

Versions

Select a version to see what changed compared to the one before it.