- In 2017, Beijing announced the goal of becoming the world’s leading AI power by 2030, and is now very close. In 2025, China is expected to spend nearly $100 billion (≈€85.7 billion) on AI, combining both state and private sectors.
- DeepSeek – an emerging Chinese startup – shocked the world by developing a large language model (LLM) competitive with ChatGPT and Grok but requiring a fraction of the cost and computational power. Meanwhile, Alibaba launched a powerful new AI model and expanded its global data centers; Tencent introduced Hunyuan-A13B, a faster, smarter model open to developers.
- Pedro Domingos, a professor at the University of Washington, commented: “China is no longer playing catch-up; it’s already on the frontline of AI.” He emphasized that the US is not as far ahead as many assume, and China was actually ahead of the US in deep learning research as early as 2010 (Baidu).
- China currently accounts for 14 out of the world’s top 20 AI models (according to the OpenCompass ranking, 10/18/2025) in reasoning, math, knowledge, and programming – with 9 being open-source models, while the US has no open-source models in this group.
- China’s strength lies in its population of over 1 billion online users, creating a “natural laboratory” for AI to learn and deploy extremely quickly. Models like DeepSeek, Qwen-3, and Kimi K2 are released for free, fostering domestic and global developer communities.
- However, China is still restricted from accessing advanced US chips, forcing it to self-develop its domestic semiconductor industry. Following Washington’s ban, Beijing retaliated by banning the import of US Micron chips and increasing investment in Cambricon Technologies, an AI chip company whose quarterly revenue increased 14-fold year-on-year.
- Experts suggest the US chip ban is counterproductive: forcing China to optimize older chips, seeking to “train smarter rather than stronger.” DeepSeek is a prime example – achieving high performance despite running on mid-range hardware.
- While the US still leads in frontier AI – models capable of reasoning, language understanding, and autonomous decision-making – China leads in practical applications and AI exports to developing countries, especially in Asia, Africa, and Europe, where Alibaba and Huawei build data centers and low-cost cloud computing platforms.
- Beijing also exports its AI governance standards based on “Chinese values,” aiming to influence how AI interprets history, culture, and truth. Domingos warns: “Whoever controls the large language models will control both the past and the future.”
- According to Robin Feldman (AI Law and Innovation Institute, UC San Francisco University), “the AI race has become a new kind of Cold War, where the nation that gains AI dominance will rule the world.”
📌 China is closing the technology gap with the US thanks to open-source AI, a massive domestic market, and the ability to optimize older hardware. China currently accounts for 14 out of the world’s top 20 AI models in reasoning, math, knowledge, and programming – with 9 being open-source models, while the US has no open-source models in this group. China leads in practical applications and AI exports to developing countries, especially in Asia, Africa, and Europe, where Alibaba and Huawei build data centers and low-cost cloud computing platforms. Beijing also exports its AI governance standards based on “Chinese values,” aiming to influence how AI interprets history, culture, and truth.
