Select Language:
While nearly double the percentage of Chinese companies have fully adopted artificial intelligence, the number that see a meaningful boost in their financial results remains very low. The primary obstacle appears to be the influence of what’s known as the Token Tax, according to a report by a leading global consulting firm.
Last year, 88% of Chinese companies that ran initial AI pilots advanced beyond trial phases, a significant jump from 46% the year before, based on the latest China Enterprise Digital Transformation Index. However, only 14% of these firms managed to convert their AI investments into increased revenue, profits, or efficiency gains—a modest rise from 9%.
The report analyzed 160 Chinese enterprises spanning eight major sectors, including advanced manufacturing, pharmaceuticals, and the automotive industry. It was released just ahead of the upcoming World AI Conference being held in Shanghai.
The lag between deploying AI solutions at scale and realizing their value seems to stem from rising hidden costs, especially the Token Tax. In fact, the study found that token consumption for identical tasks can differ by as much as 17 times due to variations in breakdown procedures, choice of tools, and model selection.
After adopting AI, many Chinese companies report their annual budgets increasing by tens of millions of yuan—equivalent to millions of US dollars—without any clear returns, according to Cao Qifeng, the lead of Data and AI for the consulting firm in China. The core issue isn’t the capability of the models themselves but whether organizations clearly understand the problems AI is meant to address and how to quantify success.
“Implementing AI internally is not just an issue of tooling; it requires strategic planning at the highest level,” Cao emphasized.
This challenge isn’t exclusive to Chinese firms. For example, Uber reportedly exhausted its entire yearly AI budget within just four months and had to introduce caps on token use among employees. Similarly, companies like Meta and Amazon have eliminated internal token consumption rankings.
Many companies’ focus over the past year on ranking employees or teams by token usage has been misguided, according to Qiu Jing, head of research in China for the firm. Nonetheless, businesses are beginning to recognize the importance of deploying different models tailored to specific tasks and scenarios, signaling a shift towards a results-driven approach.
Only 10% of Companies Are Ready for Next-Generation Competition
This year, China’s digital transformation index hit 59 points, marking the fastest growth over nine years. To assess whether firms are equipped for future competition, the consulting firm developed an evaluation framework centered on three key capabilities: autonomous intelligence, value density, and multi-polar operations.
Results show only 10% of the surveyed companies rank in the top half of their industry in digital transformation, with at least two of these future capabilities in the top 20%. While leadership’s enthusiasm for AI is increasing—61% of companies now have top management leading AI initiatives—the actual implementation remains superficial. Only 18% of the 63% that have deployed AI assistants have made substantial process redesigns.
Cao noted that successful AI adoption requires collaboration across departments and functions, not just within IT or specific business units. He likened this transformation to the shift from steam engines to electric motors in the 19th century. Factories initially installed motors without redesigning their processes, which didn’t improve efficiency. Real gains only came after comprehensive process reengineering around the new power source.
From Tool to Catalyst for Growth
Eighty-two percent of Chinese companies aim to use AI to unlock new growth opportunities rather than just cut costs. Yet, only 14% have achieved tangible value, such as a 10% or more boost in efficiency or a 5% or higher increase in revenue or profits.
Some case studies illustrate how AI can drive real transformation. For instance, Taikang Life Insurance integrated extensive underwriting rules, a comprehensive medical knowledge graph with 1.8 million entries, and insights from 50,000 historical cases into an AI system. This has resulted in underwriting accuracy exceeding 95%, enabling agents to deliver complete advice within a minute of a medical report upload.
Du Yanbin, a member of Taikang’s leadership team, pointed out that the biggest obstacle to scaling AI stems from organizational structures, processes, and skill sets that are outdated, despite AI’s emergence as a productive force.
GE Healthcare China started developing a data localization platform six years ago, establishing local governance standards that are now being adopted globally within GE Healthcare.
Facing a saturated charging network and standardized hardware, charging station operator Star Charge turned away from simply expanding pile numbers and instead adopted AI for managing complex, multi-variable dispatching involving photovoltaics, energy storage, and charging.
For example, a large industrial park using the AI system reported saving 21% on charging costs and boosting revenue by 22%.
AI as a Key to Global Growth
Increasing numbers of Chinese firms are expanding internationally to reduce domestic market pressures. According to the report, 59% of companies see tougher regulations and compliance requirements abroad, and 58% cite rising competition at home—up significantly from previous years.
China’s outbound investments grew 7.1% last year to $174.4 billion, with more than 50,000 overseas enterprises operating across 190 countries and regions.
Qiu Jing explained that AI agents are creating crucial new pathways for Chinese companies’ global expansion. As overseas consumers increasingly rely on AI-driven comparisons and recommendations, traditional legacy brands are losing some of their previously established advantages. Products and services are now directly evaluated and recommended by intelligent agents based on quality and strength.
However, even with ambitions to become global leaders—more than 60% of Chinese companies express this goal—only 5% have achieved significant global brand influence, and just 12% can shape the development of international industries.
The underlying challenge is narrative style. In other countries, companies often emphasize contributions to local ecosystems, industries, or consumers rather than just exporting products, technologies, and standards.
Looking ahead, the next step is not merely refining current growth strategies but leveraging AI to rebuild capabilities and create new value-creating models, opening up fresh growth opportunities in a diverse global landscape.
Succeeding in the AI era is less about technology and more about leadership. Companies must effectively integrate technological capabilities, business scenarios, organizational structures, and governance to transform AI from a standalone efficiency tool into a strategic engine for sustainable, high-quality growth.
ChatGPT
Perplexity
Gemini AI
Grok AI
