Industry News

Staff members work at China Mobile's computing center in Qingyang, northwest China's Gansu Province, Jan. 6, 2025. (Xinhua/Lang Bingbing)

BEIJING, May 3 (Xinhua) -- In southwest China's Guizhou Province, Tang Hanlin, chairman of a data asset operation and digital-intelligence services provider, has witnessed a dramatic surge in just four months.

His company's daily token calls skyrocketed from less than 3 billion in early January to nearly 12 billion in April, while order transactions jumped from under 70,000 to over 280,000.

This explosive growth is not an isolated case. Across China, the national daily token calls exceeded 140 trillion in March this year, representing a more than 1,000-fold surge from 100 billion at the beginning of 2024, according to official data. Tokens, the basic units that AI models use to process information, have become a barometer of industry development and a key indicator of AI adoption.

"The massive increase in daily token calls fully demonstrates that China's AI development has entered a rapid growth phase," said Liu Liehong, head of the National Data Administration. "From conversational AI to decision-making intelligent agents, China's AI industry competitiveness has significantly strengthened."

China's 2026 central government work report, for the first time, called for the creation of "new forms of a smart economy."

The smart economy's ripple effects are reshaping entire supply chains. In the first quarter of 2026, upstream electronic specialty materials manufacturing jumped 32.5 percent year on year, midstream integrated circuit manufacturing surged 49.4 percent, and downstream smart equipment manufacturing expanded 16.9 percent.

The smart economy is spawning new business models and operational paradigms as barriers to starting a business crumble amid more accessible, affordable AI tools.

In an "AI+" industrial community in Hangzhou, east China's Zhejiang Province, entrepreneur Pang Haoyang is directing a team of a dozen "AI employees" to produce AI-generated short films -- handling scriptwriting, visual generation, and post-production independently.

"With large language models, computing platforms and AI tools, one person can now do what used to require an entire team," Pang said. Since his company was founded early last year, it has already generated more than 1 million yuan (about 145,713 U.S. dollars) in revenue.

Not just confined to tech firms and internet giants, the trend is reshaping traditional industries -- a point well illustrated at Yueda Textile Group's 100,000-spindle green smart plant in Yancheng, east China's Jiangsu Province.

Rows of digital devices operate at high speed, while automated guided vehicles glide along intelligent transport routes.

"Spinning used to rely on veteran workers," said Liu Qi, the company's information center director. "Now it relies on an industrial brain that can think and optimize."

The outcomes are impressive. Since the AI-driven transformation, the plant's overall labor productivity has more than quadrupled, energy consumption per unit of output has fallen by 15 percent, and processing costs have dropped by 26 percent.

Elsewhere, Goneo Group, a manufacturer of electrical accessories, has also built an AI-powered knowledge base to streamline customer service. Zhu Kaijie, an engineer at the company, said the system has reduced response times and errors as AI usage expands.

The surging demand for tokens is creating significant opportunities for industries such as AI chips, data centers and green electricity.

Industry experts say that bringing down the effective cost per token will be at the heart of innovation in the computing industry. Achieving that will require closer alignment between computing power and applications, full-stack optimization of software and hardware, and tighter collaboration across the entire industrial chain.

With abundant data, a full industrial chain, and vast application markets, China is poised to break new ground in the smart economy during the 15th Five-Year Plan period (2026-2030), experts said.