DeepSeek, a prominent AI startup based in Hangzhou, is gearing up for an initial public offering (IPO) this year.
The company, founded in July 2023 by Liang Wenfeng, has gained significant attention since its launch of the DeepSeek-R1 reasoning model in January 2025, which featured 671 billion parameters and was trained at a remarkably low cost.
Funding and Valuation
As of June 2026, DeepSeek has not made an official IPO announcement, unlike other AI firms that have pursued confidential filings. The startup is finalizing a funding round of approximately 50 billion yuan, around $7.4 billion, which values the company between $52 billion and $59 billion. This valuation positions DeepSeek among the top AI-focused firms globally, although still below its American counterparts like OpenAI.
Backing for this funding round reportedly includes investments from major players such as Tencent and CATL, as well as a significant personal contribution from the founder. The low training cost of its flagship model, at just $5.6 million, contrasts sharply with the hundreds of millions typically needed for similar models in the West.
Geopolitical Landscape
DeepSeek's IPO plans unfold amid increasing US-China tensions concerning AI and semiconductor exports. The ongoing restrictions imposed by Washington on advanced chip sales to China have influenced the company’s decision to train its model on Nvidia’s H800, a version adapted for compliance.
Additionally, DeepSeek is branching into custom AI chip development, aiming to lessen reliance on Nvidia's GPU dominance. This move aligns with a broader trend among technology companies seeking greater independence in the evolving AI sector.
This material is informational and not financial advice.
