DeepSeek's first funding round is fiercely contested, with valuation potentially reaching $45 billion

According to Western media reports, China's National Integrated Circuit Industry Investment Fund (the "National Big Fund") is in talks to lead DeepSeek's first funding round, with the company's latest valuation estimated at around $45 billion (approximately 30.65 billion RMB), drawing significant market attention.

Sources indicate that the National Big Fund and DeepSeek are indeed in negotiations over financing, though the valuation has not yet been finalized. Multiple internet giants and other state-owned investment funds are also quietly vying for participation, and the final list of investors remains uncertain.

As a key driver in advancing China’s independent semiconductor ecosystem, the "National Big Fund" has previously invested in critical enterprises within China’s semiconductor industry, including SMIC—the largest and most technologically advanced chip foundry in China—and YMTC, China’s leading manufacturer of memory chips.

DeepSeek rapidly gained fame in January 2025 after releasing its open-source large model R1, achieving impressive results with only a fraction of the training compute power used by its U.S. counterparts. Despite not yet focusing on commercialization, its valuation has surged from $20 billion just weeks ago when fundraising began. In its latest V4 model, DeepSeek states it has optimized inference operations specifically for Huawei's Ascend 950PR chip.

Earlier foreign reports indicated that Huawei AI chips have already surpassed NVIDIA in China’s domestic market, while NVIDIA’s advanced products remain banned from export to China. However, China’s overall AI chip production capacity remains a tiny fraction of that in the United States and lags by at least two generations. To close this gap, Beijing is pushing closer collaboration between domestic chipmakers and AI model companies, aiming to build an autonomous AI ecosystem amid increasingly stringent U.S. export controls.

NVIDIA CEO Huang Renxun believes such an ecosystem could pose a threat to America’s technological dominance. In a recent interview with podcast host Dwarkesh Patel, he stated: "If one day DeepSeek launches first on Huawei’s platform, that would be a terrible outcome for our country."

He further warned that this scenario might lead to a situation where "AI models developed globally end up performing best on non-U.S. hardware."

Since entering the global spotlight last year, DeepSeek has remained focused on training cutting-edge AI models rather than building commercial businesses selling AI services to enterprises or expanding consumer-facing AI chatbot offerings.

Source: sputniknews

Original article: toutiao.com/article/1864514697098240/

Disclaimer: The views expressed in this article are solely those of the author.