China’s frontier AI race is heating up, with DeepSeek in talks to raise fresh capital at a potential valuation of around $45 billion, as the state-backed China Integrated Circuit Industry Investment Fund (Big Fund) looks to lead its first external funding round, according to a Financial Times report.
The proposed valuation marks a sharp jump from the roughly $20 billion tag DeepSeek was pegged at just weeks ago, effectively doubling as investor interest intensifies around the startup’s fast-evolving large language model (LLM) capabilities.
Apart from Big Fund, other investors, including Chinese tech giant Tencent and Flounder, are also in discussions to pick up a stake, though the final investor lineup is yet to be locked.
DeepSeek shot into the spotlight earlier this year following the launch of its R1 open-source model, positioning itself as a serious challenger to global leaders like OpenAI and Anthropic, despite operating with significantly lower compute resources. The company, led by billionaire founder Liang Wenfeng, has largely focused on building frontier models rather than aggressive commercialization.
For Big Fund, the potential bet on DeepSeek aligns with Beijing’s broader push for semiconductor and AI self-reliance. The fund, which has raised about $47 billion in its latest phase, has historically backed key chipmakers such as Semiconductor Manufacturing International Corporation and Yangtze Memory Technologies Corp., but has not previously invested in LLM developers.
A DeepSeek investment would signal a tighter integration between China’s chip ecosystem and AI model builders, especially as the startup optimizes its models to run on domestic hardware such as Huawei’s Ascend chips, part of a broader effort to reduce reliance on US technology.
Within China, DeepSeek is emerging among a cohort of highly valued AI players, alongside companies like Zhipu AI, which commands a market value of about $52 billion, as well as other fast-scaling labs building frontier models.
Globally, the opportunity is massive. As per statista, the large language models market was valued at $5.6 billion in 2024 and is projected to grow at a CAGR of 36.9% between 2025 and 2030, underlining why capital is rapidly flowing into the space.
