China is stepping up its game in artificial intelligence (AI) as it faces tough U.S. restrictions on advanced technology. At the recent World Artificial Intelligence Conference (WAIC) in Shanghai, ending July 28, 2025, Chinese AI companies announced two new alliances to create a powerful domestic AI ecosystem. These moves aim to reduce reliance on foreign tech, like Nvidia’s advanced chipsets, and boost China’s global AI influence.

U.S. has placed strict export controls on high-end AI chips, limiting Chinese companies’ access to cutting-edge technology. To overcome this, China is focusing on building its own AI systems. The two new alliances, announced at WAIC, bring together Chinese AI companies to work on everything from chips to large language models (LLMs) and infrastructure. This teamwork helps China create self-sufficient AI technology and compete globally.

The Model-Chip Ecosystem Innovation Alliance, led by StepFun, connects Chinese developers of large language models with chipmakers like Huawei, Biren, and Moore Threads. These companies have been hit by U.S. sanctions but are now collaborating to build a complete AI technology chain.

“This alliance creates a strong ecosystem, linking chips, models, and infrastructure,” said Zhao Lidong, CEO of Enflame, a chipmaker in the alliance.

The second alliance, the Shanghai General Chamber of Commerce AI Committee, includes companies like SenseTime, StepFun, MiniMax, Metax, and Iluvatar CoreX. This group focuses on using AI to transform industries, such as manufacturing and healthcare, in China and beyond.

The WAIC in Shanghai was a hotspot for new AI products. Here are some highlights:

• Huawei’s CloudMatrix 384: This AI computing system uses 384 of Huawei’s 910C chips and rivals Nvidia’s top systems, according to SemiAnalysis. It shows China’s ability to innovate despite U.S. restrictions.
• Metax’s AI Supernode: This system uses 128 C550 chips for large-scale, liquid-cooled data centers, perfect for advanced AI tasks.
• Tencent’s Hunyuan3D World Model 1.0: An open-source tool that creates interactive 3D environments from text or images.
• Baidu’s Digital Human Technology: This creates virtual livestreamers by copying a person’s voice and movements from just 10 minutes of video.
• Alibaba’s Quark AI Glasses: Powered by the Qwen AI model, these glasses help users navigate, shop, and use Alipay with voice commands. They will launch in China by late 2025.

China is betting big on open-source AI models to gain global influence. Companies like DeepSeek and Moonshot AI have released powerful models, R1 and Kimi K2, which compete with top U.S. models. These models are free to customize, making them popular in developing countries. For example, DeepSeek’s R1 had 97 million active users worldwide by April 2025, with over 500 customized versions downloaded 2.5 million times in January alone.

This open-model approach allows countries in Africa, Asia, and Latin America to adapt AI for local needs, like healthcare or education, at a low cost. Chinese Premier Li Qiang emphasized this at WAIC, saying, “Openness and sharing in AI can benefit more countries.” This strategy boosts China’s soft power, helping it build goodwill and influence globally.

The U.S. and China are in a heated race for AI dominance. While the U.S. focuses on closed AI models (like OpenAI’s ChatGPT), which are powerful but expensive, China’s open models are cheaper and more accessible. U.S. export controls limit China’s access to advanced chips, but Chinese firms are finding ways to innovate, like Huawei’s chip-clustering technology.

However, the U.S. is responding. OpenAI plans to release an open-source model in late 2025, and Google has launched its Gemma models. The Trump administration’s AI Action Plan, released in July 2025, aims to boost U.S. AI exports and promote open models based on “American values.”

At WAIC, Premier Li Qiang called for a global consensus on balancing AI development and security. He announced a Chinese-led Global AI Governance Initiative to promote responsible AI use and cooperation, especially with developing countries. China wants to create an international AI organization based in Shanghai to share its AI advancements and set global standards.

This move counters U.S. efforts to limit China’s AI growth. Li warned against “technological monopolies” that make AI exclusive to a few countries. He highlighted challenges like limited chip supplies and the need for fair AI access worldwide.

China’s AI alliances and open-source strategy could reshape the global tech landscape. By offering affordable, customizable AI models, China is gaining influence in the Global South (developing countries in Africa, Asia, and Latin America). This could challenge U.S. leadership in AI and shift global tech influence.

For the U.S., balancing security with innovation is key. While export controls slow China’s progress, they haven’t stopped it. The U.S. must invest in open AI models to stay competitive and maintain its global influence.