DeepSeek pauses fundraising talks with prospective investors: Report
DeepSeek was seeking fresh capital at an estimated $71 billion pre-money valuation before the suspension.
DeepSeek has temporarily suspended its latest fundraising talks after telling some prospective investors that planned investment agreements would not be finalized in the coming days, Bloomberg reported Saturday, citing people familiar with the matter.
The decision follows the viral spread of comments attributed to founder Liang Wenfeng regarding China’s AI capabilities and reliance on Nvidia chips. Sources said Liang was dissatisfied with reports describing remarks allegedly made during investor meetings for DeepSeek’s June funding round, which raised $7 billion.
The AI startup had been pursuing at least 10 billion yuan ($1.5 billion) in new funding at a pre-money valuation of roughly 480 billion yuan ($71 billion), just weeks after completing a record $7 billion financing led by investors including Tencent, CATL and a state-backed AI investment fund.
Although fundraising has been put on hold, discussions remain ongoing and DeepSeek could restart the process.
The company is also laying the groundwork for an IPO that could be filed later this year as it seeks capital to expand computing capacity while continuing to focus on open-source AI development and long-term artificial general intelligence research.
DeepSeek founder says China trails US because of computing resources
DeepSeek believes computing power, not talent, is the primary factor separating China’s AI industry from the US, according to unconfirmed investor meeting minutes reported by Yicai.
Wenfeng reportedly said restrictions on advanced Nvidia AI chips and lower capital investment have constrained China’s AI development, although the company estimates that it remains only 12 to 18 months behind leading US firms while using around one-twentieth of their computing resources.
The reported meeting also said DeepSeek has computing resources equivalent to approximately 20,000 Nvidia H-series GPUs and plans to shrink the technology gap to three to six months.
The startup is said to be working with Huawei to optimize its models for Ascend chips and reduce its dependence on Nvidia’s CUDA software ecosystem.