With 87 percent of Chinese companies planning to increase investment in artificial intelligence in 2025, and more than half reporting faster-than-expected progress, China's AI momentum is accelerating. At the same time, the AI+ Action Plan of the State Council, China's Cabinet, is pushing adoption across priority sectors, from manufacturing and energy to healthcare, finance and retail. Yet, many companies still face a common bottleneck: moving from promising pilot projects to scalable, measurable productivity gains.

At the World Economic Forum in Davos, Switzerland, business leaders and technology executives pointed to a distinctly Chinese pathway for AI development, one that emphasizes applications, industrial integration and cost-effective deployment, rather than a narrow focus on frontier models alone.

Dora Liu, CEO of Deloitte China, described the global AI race as a long-term competition shaped by infrastructure and real-world demand. "The AI race is definitely not a sprint. It is a marathon," she said, adding that over time, competition will hinge on energy, data centers and electricity supply, areas where China has strong foundations.

While acknowledging that China still trails the United States in advanced chips, Liu stressed that AI competitiveness ultimately depends on where and how technology is applied. China's economy provides abundant scenarios, ranging from manufacturing all the way to new medicine development and the service industry, she said.

That application-first approach was echoed by Joe Ngai, chairman of McKinsey Greater China, who said Chinese companies are focusing less on debating model architectures and more on improving productivity. "Chinese companies are actually very practical, they're using AI to think about applications," he said, from helping employees "do more with less" to deploying what he called "embodied AI".

Ngai highlighted the combination of AI with China's manufacturing and supply-chain ecosystems, which enables smarter automation across factories and logistics networks. "When you combine the supply chain with AI, then you have all these automations that happen that are very smart," he said, describing China's trajectory as one of "applied AI or physical AI", where software intelligence is tightly coupled with hardware systems.

On the consumer side, AI is increasingly shaping purchasing behavior and product design. Xu Ran, CEO of JD.com, said 2025 marked a turning point for AI-related consumption on the platform, which now serves around 700 million annual active customers. "2025 was a breakout year for AI products," she said, noting that sales of products labeled as "smart" doubled year-on-year, while devices with stronger AI capabilities, such as robots and smart glasses, saw sales grow by three times and 10 times, respectively.

Search behavior revealed even stronger signals of interest. Xu said search volume for AI-related products "skyrocketed by 100 times" in 2025, and customer surveys showed that nearly half of users now consider AI a "must-have function" rather than a premium add-on.

"AI function is no longer a nice-to-have feature; it's a must-have core product value to drive consumption and growth in the retail industry," she said, adding that JD is working with business partners to codevelop products that better match evolving demand.

For AI developers, China's industrial breadth offers fertile ground for scaling technologies beyond laboratory settings. Zhang Yutong, president of startup Moonshot AI, said the country's vast manufacturing and retail sectors create an environment where AI systems can be tested and deployed at a real scale. "There will be much data, many use cases that people can try, use AI to couple with," she said, calling this one of the market's major advantages.

Zhang also pointed to China's openness to new technologies, citing electric vehicles, solar energy, smartphones and autonomous driving as examples of rapid consumer acceptance. She said she was personally surprised that about 85 percent of Chinese respondents consider autonomous driving safe and beneficial for productivity, helping explain why thousands of robotaxis are already operating in dozens of cities.

Enterprise deployment, however, still hinges on cost and return on investment. Dowson Tong, CEO of Tencent Cloud and Smart Industries Group, emphasized that AI in practice consists of many specialized models serving different business functions, rather than a single monolithic system. Across industries, he said, companies are using AI "in every single function within their operation".

Tong said retailers are using generative photo and 3D modeling tools to shorten design cycles, while marketing teams rely on AI for better targeting and personalization to improve conversion rates.