Chinese AI startup Deepseek is close to raising at least $12 billion in a new funding round that could reach up to $15 billion, according to a report from Bloomberg. The round is heavily backed by battery maker CATL and tech giant Tencent. The capital injection follows investor enthusiasm for Deepseek’s V4-Flash model, which achieved strong cost and performance benchmarks compared to models from OpenAI and Anthropic.

Following the round’s completion, Deepseek plans to restructure its corporate setup to prepare for an early 2027 initial public offering. The startup is currently constructing a specialized data center powered by at least 160,000 Huawei AI chips. Founder Liang Wenfeng maintains that the company will remain committed to developing open-weight AI models despite commercial scale.

The massive round follows a $7.4 billion capital raise in June that valued the company above $50 billion. Deepseek previously paused fundraising after viral remarks regarding its reliance on Nvidia hardware. To mitigate supply dependencies, Deepseek is actively designing its own custom inference chips alongside its deployment of Huawei hardware.

Why it matters

  • Founders using open-weight models gain long-term stability as Deepseek secures massive capital for open models.

  • Hardware operators should track Deepseek’s custom inference chip and Huawei deployment as Nvidia alternatives.

  • Investors face escalating late-stage capital requirements as private AI rounds approach $15 billion scales.

Source: the-decoder.com