Crouching tigers, hidden giants narrow China’s AI gap with US

In 2017, China published the Next Generation AI Development Plan, aiming to become the global leader in AI technologies by 2030. Since then, funding for Chinese AI startups has risen sharply.

Last year, the segment made up 16% of total funding in China from just 6% in 2023, according to Tech in Asia’s database. 

China has also accelerated the launch of notable AI models, according to Epoch AI’s data. The research institute categorizes leading AI models as one that meets any of the following criteria: is a state-of-the-art improvement on a recognized benchmark or has over 1,000 citations, historical relevance, or significant use.

Last year, Chinese organizations developed 24 leading AI models — nearly half the US total — marking the smallest gap between the two countries since 2010.

As of April 23, two of the 10 best-performing large language models (LLMs) globally are from Chinese firms, either developed by Alibaba or DeepSeek. 

This also represents the two groups in China’s AI market: the tech giants and the rising startups.

Chinese tech giants like Alibaba, Baidu, Tencent, and ByteDance are heavily investing in AI, which is now seen as the most crucial technology. These firms have all developed their own LLMs since 2015 and have declared billions of dollars in investments for AI in the coming years. 

Besides this, startups are also fueling China’s AI landscape. The most well-known name is DeepSeek, which has stunned US AI giants such as OpenAI, Meta, and Google with its cost-effective open-source R1 model launched in January. 

In addition, China has six “AI tigers,” a term used by local media to describe generative AI startups that have reached billion-dollar valuations. These are Zhipu AI, Minimax, Moonshot AI, StepFun, 01.AI, and Baichuan AI. 

That said, tech giants and AI startups in China don’t compete directly. Apart from DeepSeek, which is owned by Chinese hedge fund High-Flyer, the six “AI tigers” are all backed by major Chinese tech firms. 

Some of them, including Baichuan, Moonshot, Zhipu AI, and StepFun, also share another common trait: receiving funding from state-linked investors. 

For many years, China has recruited Chinese science and tech experts educated overseas and have lured them to return home. But in the AI sector, the country’s education system has proved its ability to nurture local talent. 

Looking at the educational background of key figures from leading Chinese AI startups, seven out of 10 earned their highest degree in China, and only three were educated in the US. Especially at DeepSeek, all core team members including scientists and developers attended local universities. 

Even though China is behind the US in the AI talent race, it’s quickly catching up.

In 2019, China didn’t even appear on the list of top countries employing the leading 2% of AI researchers. But three years later, it surged to second place with 12%, according to research by MacroPolo, the former think tank of Paulson Institute.