The $7bn opportunity: Why DeepSeek still dominates the AI market

DeepSeek, the Chinese AI startup that dominated the headlines last year, has raised more than $7.4 billion (50 billion yuan) in its first funding round.

First reported by the Information, the funding round means DeepSeek is more valued at more than $50 billion. The financing structure required investors to put their capital into a limited partnership managed by CEO Liang Wenfeng, according to the Information sources, which enables him to retain absolute control over the company.

A five-year lockup has also reportedly been imposed on all investor shares, with investors unable to sell their stakes during this time. The exception is China’s National Artificial Intelligence Industry Investment Fund, which has invested directly in DeepSeek and has retained voting rights and freedom from the lock-up, according to the Information.

Reuters has previously reported that Liang committed 20 billion yuan of his own money into this funding round and suggested that Tencent had considered 10 billion yuan and CATL had considered five billion yuan. If this is the case, this would make them the largest external investors, Reuters said.

The DeepSeek funding round attracted notable interest from significant investors in China, including Tencent and CATL. The AI startup is still strongly regarded as one of the leading AI developers of China, having garnered international attention with its low operating costs and advanced capabilities that are said to rival OpenAI’s ChatGPT.

What gives DeepSeek its edge in the market is the AI’s ability to reason step-by-step. It was only the second firm, after OpenAI, to release a reasoning model – mimicking human reasoning by breaking tasks into intuitive steps and giving feedback at every stage.

DeepSeek has since become a national AI champion for China, having challenged US assumptions about its technology capabilities last year. The general belief was that China was far behind the US in LLM technology, with DeepSeek acting as a large wake-up call.

The release of its V4 model earlier this year triggered a scramble for chips, with China’s largest technology companies – including ByteDance, Tencent and Alibaba all vying for Huawei’s Ascend 950 AI chips, which DeepSeek had optimised V4 for, instead of opting for Nvidia hardware.

With US export restrictions continuing to limit Chinese firms’ access to Nvidia’s most advanced processors, domestic alternatives have become the core strategy for many Chinese cloud and AI operators.

Big picture, DeepSeek’s ability to scale the global AI market could help speed up the ongoing shift from training to inference for LLMs. The change is happening as data centre hyperscalers continue to spend billions on AI infrastructure, with the likes of Microsoft, Meta and SoftBank leading the charge.

China’s Big Fund in talks to lead DeepSeek investment at $45bn valuation

Huawei targets 60% jump in AI chip revenue as DeepSeek drives domestic surge