DeepSeek has told potential investors it is pausing agreements for its second funding round, according to Bloomberg, after remarks by founder Liang Wenfeng on US-China AI competition drew online attention and left management frustrated enough to halt the process.
Second fundraising round put on hold
Bloomberg reported that the Chinese artificial intelligence startup completed a $7 billion first-round financing in June 2026 and then moved into a new round targeting at least RMB 10 billion. The company was seeking a pre-money valuation of more than RMB 480 billion.
People familiar with the matter said Liang recently spoke at an internal meeting about China’s reliance on Nvidia chips for AI development and about how Chinese technology still lagged behind the United States. Records of that meeting later circulated online and were widely picked up by media outlets.
After that episode, management orally informed some potential investors that financing agreements originally expected to be signed soon would be paused, citing concerns over the leak of non-public comments. Bloomberg added that the negotiations remain subject to change and the fundraising process could still restart.
RMB 480 billion valuation target and IPO planning
In its first fundraising round in June, DeepSeek raised $7 billion, bringing its valuation to about $50 billion. The second round that followed was aimed at raising at least RMB 10 billion, with a pre-money valuation set at more than RMB 480 billion.
Alongside private fundraising, the company has also begun planning for an initial public offering. Bloomberg said it could file for an IPO as early as the end of this year.
Backers and research priorities
DeepSeek was founded in 2023. Its parent company is quantitative hedge fund Zhejiang High-Flyer Quant. Current shareholders include Tencent Holdings Ltd., CATL, and the state-backed National AI Industry Investment Fund.
The company’s leadership has told investors that its operating strategy will prioritize breakthrough AI research over short-term commercial monetization. Liang has also pledged to continue building open-source models, with artificial general intelligence, or AGI, as the long-term goal.
Low-cost model development and Liang’s net worth
Bloomberg said DeepSeek’s high-performance model released in 2025 shook the industry by showing that Chinese companies could still build top-tier AI platforms with fewer computing resources even under US restrictions on advanced hardware exports.
After the first financing round closed, Liang’s net worth rose from $16.7 billion to $36 billion. According to the Bloomberg Billionaires Index, that put him ahead of the founders of US rivals including OpenAI and Anthropic, making him the world’s richest AI model maker.