Hangzhou, June 17, 2026
The Chinese AI lab DeepSeek has raised more than 50 billion yuan (around $7.4 billion) in its first external funding round and, according to consistent reports, is being valued at over $50 billion.
What is new as of June 17, 2026
Update of June 17, 2026: The Chinese AI lab DeepSeek has closed its first external funding round, generating more than $7.4 billion according to the technology portal "The Information." The valuation of the Hangzhou-based company thus rises to over $50 billion – making DeepSeek, according to Chinese media reports, the country's most valuable AI startup. The round marks a turning point for the lab, which until now had been financed almost exclusively from private funds of the founder.
The bulk of the fresh capital is contributed by DeepSeek founder Liang Wenfeng himself: around 20 billion yuan (about $3 billion) comes from his private assets, according to people familiar with the matter. This means the 2023-founded quant hedge fund supplier High-Flyer remains the lab's dominant source of funding. DeepSeek was founded in July 2023, when Liang provided the necessary start-up capital from his hedge fund High-Flyer and from private funds.
Background: The role of the founder
Among the few external investors are two prominent names from Chinese industry: the technology conglomerate Tencent and the battery manufacturer CATL. Tencent invested around 10 billion yuan (about $1.4 billion) according to the reports, while CATL contributed around 5 billion yuan. However, Tencent's check grants the conglomerate no governance influence whatsoever – the company has no say in the new structure.
An exception is a Chinese state investment fund, which as the only external investor was granted a seat on the control board. Which fund is specifically involved and what exact conditions apply does not emerge from the available reports. The news agency Reuters had previously already reported on the founder's personal participation.
International comparison
Internationally, DeepSeek's capital raise remains modest: US competitor OpenAI most recently closed a funding round of $122 billion at a valuation of around $852 billion, while Anthropic raised $65 billion. DeepSeek, at $7.4 billion, sits well below these figures – even if its valuation represents an enormous jump in percentage terms relative to the industry leader.
What is new as of June 17, 2026: With the closing of the round, a concrete valuation figure is available for the first time. Previously, it had only been known that DeepSeek wanted to take in external funds. Now, both the volume of $7.4 billion and the valuation of over $50 billion have been made public. Also new is the confirmation that Liang Wenfeng alone is contributing 20 billion yuan out of his own pocket – and thereby holding by far the largest share of the round. The explicit mention that Tencent, despite its billion-dollar investment, receives no governance influence is also new information. Until now, there had only been speculation about the form in which Chinese tech conglomerates would be involved.
The distillation accusation from Anthropic
In parallel to the funding news, DeepSeek has been facing a serious accusation since February 2026: the US provider Anthropic, developer of the Claude model, accused DeepSeek – together with the Chinese labs Moonshot AI and MiniMax – in a blog post of having extracted capabilities from Claude on an "industrial scale." At the center is the concept of distillation: a process in which a smaller "student" model is trained on the outputs of a stronger "teacher" model.
Specifically, the three companies are said to have generated more than 16 million interactions with Claude through around 24,000 fraudulently created accounts – a violation of the terms of service and the regional access restrictions, since Claude is not commercially available in China. Anthropic attributed the bulk of the data traffic to MiniMax (over 13 million interactions); DeepSeek is said to have carried out around 150,000 targeted exchanges aimed at Claude's reasoning capabilities and at censorship-compliant answers on sensitive topics.
DeepSeek has not publicly commented on the accusations so far. There is also no direct statement from Liang Wenfeng on the funding round. The round apparently proceeded without the participation of Western capital providers, which is hardly surprising given the geopolitical tensions in the semiconductor and AI sectors.
Market position of DeepSeek V4 Pro
What the round means for competition in the AI sector also depends on the market position of DeepSeek V4 Pro. On the Chatbot Arena (LMArena), which is based on millions of blinded user comparisons, the leading models move within an Elo range of around 1,450 to 1,560; DeepSeek's reasoning models line up at the lower end of this frontier range and score particularly well in the categories of mathematics and logic. On specialized comparison lists of Chinese models, DeepSeek V4 Pro leads competitors such as GLM-5.1, Kimi K2.6 and Qwen.
In the Intelligence Index from Artificial Analysis, which combines reasoning, speed and cost, DeepSeek V4 Pro sits well above the median of comparable open-weight models, but in the absolute top tier it remains behind the leading proprietary systems from Anthropic, OpenAI and Google. Its central selling point is price: at around $0.45 per million input tokens, DeepSeek V4 Pro is among the cheapest models in the near-frontier segment and undercuts the top providers by roughly an order of magnitude.
This combination of competitive performance and low price had already propelled DeepSeek into headlines worldwide in early 2025 with the release of a chatbot. Since then, the lab has been regarded as a significant force in the competition between Chinese and Western AI providers – and not least as an argument that powerful AI models do not necessarily require hundreds of billions of dollars in computing infrastructure.
Outlook: What happens with the fresh capital?
With the fresh capital, DeepSeek is likely to invest primarily in three directions: in the expansion of training infrastructure, in attracting additional AI researchers, and in the international marketing of the open-weight model. It remains unclear whether the company plans to license the model more strongly outside China or whether the open-weight strategy will be maintained.
The valuation of over $50 billion also sets a new benchmark for the entire Chinese AI startup scene. Industry observers see the round as a signal that Chinese venture capital investors are willing to build local champions without interfering in entrepreneurial leadership. Whether this structure holds in the long term will also depend on how DeepSeek deals with the ongoing distillation accusations and whether Western regulatory authorities – in particular the EU Commission – take measures that further restrict market access for Chinese models.