The founder of DeepSeek announces that achieving AGI is the objective as the $10 billion funding round progresses.

      Liang Wenfeng has informed potential investors that the Hangzhou lab will focus on cutting-edge research rather than immediate revenue in its inaugural external funding round. The founder of DeepSeek stated that the lab aims to target artificial general intelligence as a major objective and will continue to release open-source models instead of pursuing short-term commercialization.

      Bloomberg reported this communication on Friday, in the context of an ongoing fundraising effort estimated at 70 billion yuan (approximately $10 billion). This amount reflects the valuation that DeepSeek is aiming for, not the size of individual investments. The startup is looking to secure a minimum of $300 million in external funding during this round. Additionally, The Information has reported that the long-term goal could escalate to $7 billion or more as the lab progresses toward generating recurring revenue.

      This round marks Liang's first acceptance of external funding. This development is significant news since DeepSeek had been exclusively financed by High-Flyer Quant, the quantitative trading firm Liang established, which continues to support the lab. Liang has previously described the lack of outside investors as a strategic choice to avoid pressure regarding the product roadmap.

      The change in approach for this round is partially due to the scale of operations: training at the dimensions DeepSeek currently employs has surpassed what even a successful hedge fund can independently finance. Additionally, the focus on AGI indicates the company's self-positioning. In April, DeepSeek launched its V4-Pro and V4-Flash models, featuring a 1.6-trillion-parameter Mixture-of-Experts system and a smaller 284-billion-parameter variant, both released open-source under lenient licensing terms.

      The V4 series is optimized for use on Huawei Ascend and Cambricon silicon, as well as Nvidia, signaling a strategic approach to a domestic Chinese market increasingly isolated from high-end US accelerators. The pitch to investors, according to Bloomberg, emphasizes the research-oriented identity that led to the January 2025 release of the R1 model. The launch of R1 reportedly led to a nearly $600 billion drop in Nvidia's market capitalization in just one trading session, based on the perception that frontier reasoning models could be developed at a fraction of the costs incurred by US laboratories.

      The validity of that pricing claim sparked its own debate, but the strategic point remained clear: a Chinese lab could operate at the forefront and do so transparently. The details that have yet to be disclosed are the most critical; investor identities have not been confirmed, the company has not formally announced the valuation, and the closing date is still undecided.

      Historically, Deep Seek has refrained from engaging with media inquiries, and the lab did not provide any on-the-record comments for Bloomberg’s report. A fundraising effort of this magnitude, at this valuation, with these stated objectives, is likely to attract interest from Chinese regulators, who have been deliberating the nature of state oversight for foundation-model developers over the past two years. What Liang has openly committed to is a forward trajectory: open-source, pioneering research with AGI as the framework instead of focusing on enterprise revenue. Once concluded, this round will be the first occurrence of external capital agreeing to these terms.