Insider Brief
- What Are China’s AI Tigers — China’s AI industry is witnessing a shift from facial recognition-focused firms to generative AI start-ups, with Baichuan AI, Zhipu AI, Moonshot AI, and MiniMax emerging as the new leaders.
- These AI start-ups, collectively known as China’s “four new AI tigers,” have secured major investments and are developing large language models (LLMs) to compete with Western firms like OpenAI.
- Despite regulatory hurdles and U.S. export controls on AI chips, China’s AI models are rapidly advancing, supported by government funding, foreign investment, and a growing domestic semiconductor ecosystem.
China’s artificial intelligence (AI) industry has entered a new era, with a wave of start-ups emerging as the country’s best hope of rivaling OpenAI’s ChatGPT and other Western AI models. These companies — Baichuan AI, Zhipu AI, Moonshot AI, and MiniMax — are collectively known as China’s “four new AI tigers.” As reported in the South China Morning Post, these firms have attracted significant investor attention and are reshaping the landscape of generative AI in China.
The Rise of China’s AI’s Tigers
Since OpenAI debuted ChatGPT in late 2022, China’s technology giants and AI start-ups have raced to develop their own large language models (LLMs). According to Chinese government figures, nearly 200 LLMs have been launched in the country. Among the most prominent players are the four new AI tigers, which have emerged as unicorns — start-ups valued at over $1 billion.
- Baichuan AI: Founded in 2023 by Wang Xiaochuan, the creator of Sogou, Baichuan has quickly become a major player in China’s AI sector. While the company has not disclosed the exact amount of its latest fundraising round, estimates suggest a valuation of approximately $1.8 billion. Baichuan has released three versions of its AI model, claiming that its latest iteration surpasses OpenAI’s GPT-4 in Chinese language proficiency.
- Zhipu AI: A spin-off from Tsinghua University’s Knowledge Engineering Group, Zhipu AI was founded in 2019. By the end of 2023, it had raised over $347 million from investors such as Alibaba, Tencent, and state-backed funds. It is currently valued at over $1.3 billion and is focused on developing foundational AI models.
- Moonshot AI: Founded in 2023, Moonshot AI has quickly gained traction in China’s AI race. Its chatbot, Kimi, can process up to 2 million Chinese characters in a single prompt. The company raised $1 billion in February 2024, bringing its valuation to around $2.5 billion.
- MiniMax: Based in Shanghai, MiniMax was founded in 2021 and has also surged in valuation, now estimated at approximately $2.5 billion. Reports indicate that it is currently seeking to raise at least $600 million in a new funding round led by Alibaba.
How Do These Start-Ups Compare Globally?
According to the Information Technology and Innovation Foundation (ITIF), China has become a global leader in AI research publications and is closing the performance gap with U.S. AI models. While U.S. AI firms have historically dominated AI research and commercialization, China is making rapid progress, particularly in generative AI and bilingual LLM benchmarks.
China’s AI research output is substantial, but its impact lags behind the U.S., as measured by citations and private-sector engagement, according to ITIF. Despite this, China’s AI start-ups have benefited from strong academic ties, particularly with Tsinghua University, which has served as the launchpad for many of these companies. The country is also seeing increased foreign investment in its generative AI sector, with Saudi Arabia’s Aramco leading recent funding efforts.
The Shift from “Old AI Dragons” to “New AI Tigers”
The rise of the four new AI tigers marks a shift from China’s previous focus on AI applications like facial recognition. The earlier generation of AI companies — SenseTime, Megvii, CloudWalk Technology, and Yitu Technology — focused on image processing and security applications, earning them the nickname “AI dragons.” The new tigers, in contrast, are concentrating on generative AI and LLMs, aiming to compete directly with OpenAI, Google DeepMind, and Anthropic.
Government Support and Regulatory Challenges
China’s AI sector operates under a unique financial and regulatory environment. The Chinese government has been heavily involved in funding AI initiatives, filling gaps where private venture capital lags. According to ITIF, state-directed capital funds and financial aid are proving effective in fostering innovation, especially in regions that get passed over by private capital.
However, China’s AI start-ups also face significant regulatory constraints. All AI models must be approved by China’s Cyberspace Administration before they can be publicly released. Additionally, U.S. export controls on high-end AI chips have limited China’s access to cutting-edge hardware, though companies like Huawei are developing domestic alternatives such as the Ascend 910B AI chip.
China is introducing incentives to produce chips domestically.
The analysts write: “Beijing city authorities have introduced subsidies for firms buying domestically produced AI chips to boost China’s semiconductor industry and lessen dependency on foreign technology. The initiative will provide financial support to companies based on a percentage of their investment in domestically controlled graphics processing unit (GPU) chips. In addition, at least 16 local governments, including the largest, Shanghai, are providing companies with vouchers to access subsidized processing power from large state-operated data centers that consolidate limited supplies of advanced chips.”
How Competitive Are China’s AI Models?
Despite restrictions, Chinese AI models are making impressive strides. Some open-source models, including Alibaba’s Qwen 1.5 and Zhipu AI’s ChatGLM3, have outperformed U.S. counterparts in certain benchmarks. According to SuperCLUE, a Chinese AI benchmarking platform, the performance gap between leading Chinese models and OpenAI’s GPT-4 is narrowing.
Moreover, China’s open-source LLM ecosystem is rapidly advancing. Companies like 01.AI and Baichuan AI have developed competitive models, gaining recognition for their efficiency and cost-effectiveness. The open-source nature of many Chinese models is also allowing for broader adoption and innovation across different industries.
China’s AI Tigers And The Global AI Race
China’s AI industry remains a top concern for U.S. policymakers. The United States leads in commercial AI applications, with companies like OpenAI, Microsoft, and Google driving much of the industry’s growth. However, China’s government-backed AI initiatives and its universities’ deep involvement in AI research have made it a formidable challenger.
While the U.S. still attracts the majority of top AI talent, more Chinese researchers are choosing to stay in China, strengthening its domestic AI ecosystem. Additionally, Chinese companies are leveraging extensive datasets, government-backed infrastructure, and state incentives to accelerate AI development.
China’s four new AI tigers — Baichuan AI, Zhipu AI, Moonshot AI, and MiniMax — are leaders in both the country’s generative AI revolution and the global AI race. Backed by strong academic institutions, government funding, and growing foreign investment, these start-ups are closing the gap with their U.S. counterparts. However, they still face challenges, including regulatory oversight and access to advanced semiconductor technology.
