DeepSeek, the Hangzhou-based artificial intelligence start-up, has closed its first-ever external funding round, raising approximately 50 billion yuan (US$7.4 billion) in a Series A that values the company at around 400 billion yuan (US$59.2 billion), according to sources with knowledge of the matter. The deal marks a pivotal moment for one of China's most closely watched AI companies and offers a rare window into its strategic direction.

Liang Wenfeng leads with a personal $3bn bet

In an unusual arrangement, DeepSeek founder and CEO Liang Wenfeng emerged as the single largest backer in the round, committing approximately 20 billion yuan of his own capital — nearly half the total raised. Earlier speculation had pointed to the China Integrated Circuit Industry Investment Fund, commonly known as the Big Fund, as the likely lead investor, but that did not materialise.

The move is widely seen as a deliberate effort by Liang to retain strategic control of the firm even as it opens up to external shareholders. According to a source familiar with the matter, Liang remains firmly committed to pursuing artificial general intelligence (AGI) — the point at which AI matches human cognitive capabilities — a vision he reportedly reinforced to investors repeatedly during an online meeting last month.

Why it matters: valuation and competitive context

The post-investment valuation of US$59.2 billion makes DeepSeek one of China's most valuable AI start-ups, surpassing the US$30 billion valuation of Moonshot AI and the US$17.7 billion market capitalisation of Hong Kong-listed MiniMax AI as of Wednesday's close. It still trails Zhipu AI, which carries a US$95 billion market value.

The funding landscape for Chinese AI is intensifying, with domestic start-ups racing to establish scale and technical credibility against both homegrown rivals and global players including OpenAI and Anthropic. DeepSeek's decision to raise external capital for the first time signals that even the most capital-disciplined players in the space now see a need to accelerate.

AGI focus: what DeepSeek told investors

Liang told investors at last month's meeting that what 'DeepSeek wants to do is anything relevant to enhance the intelligence' of AI models, and that it had no interest in anything else, according to a source present. The statement underscores a narrow, research-first mandate that sets DeepSeek apart from rivals that have moved aggressively into enterprise software, consumer applications, and hardware.

The company's roots in quantitative finance — Liang co-founded high-frequency trading firm High-Flyer Quant — have historically given it a distinctive approach to AI model development, prioritising efficiency and compute optimisation over brute-force scaling.

What's next

With fresh capital secured and founder control intact, the central question is how DeepSeek will deploy the funds while remaining insulated from commercial pressures that typically accompany institutional investment. The company's next model releases and any moves toward productisation will be closely watched by investors and competitors alike, including technology giants Tencent, JD.com, and NetEase, all of which are deepening their own AI commitments. How DeepSeek balances its AGI ambitions against the expectations of new external shareholders may define its trajectory for years to come.