AI has transitioned from a niche focus to a thriving ecosystem. As this transformation unfolds, more entities are redirecting their focus towards implementation and commercial deployment. Both the foundational infrastructure of AI large models and the commercial applications derived from them are undergoing a significant upgrade. The rising utilization of AI large model tokens, termed as 'ciyuan' in Chinese, serves as a compelling indicator of this shift.

On April 6, leveraging the latest data from OpenRouter, a globally recognized API (Application Programming Interface) call platform for large models, the total global AI large model token usage soared to 27 trillion tokens in the preceding week (March 30 to April 5), marking an 18.9% month-on-month (MoM) increase. China's weekly AI large model token usage surged to 12.96 trillion tokens, up 31.48% MoM, while the US's weekly usage reached 3.03 trillion tokens, growing by 0.76% MoM. China's weekly AI large model token usage has outstripped the US for five consecutive weeks, indicating sustained growth momentum.

The accelerated growth in token usage is merely one facet, while the burgeoning wave of commercialization in the Chinese market directly mirrors this broader trend. AI large models and intelligent agents are becoming ubiquitous features for an increasing number of entities. From DeepSeek to Yuanbao, Doubao, and subsequently OpenClaw, AI is seamlessly integrating into people's daily production and lives. As this trend accelerates, the demand for foundational tokens—ciyuan—is also escalating. Behind this lies a more profound transformation: the AI industry is undergoing an accelerated evolution.

I

Reflecting on the development of AI, it was initially propelled by internet players seeking new avenues amid stagnating traffic and growth challenges. In many instances, AI deployment represented their endeavors to transform traditional businesses and models. Whether applied to content creation or e-commerce, AI was primarily utilized to revamp existing industries and explore new growth avenues.

After internet players revitalized their traditional businesses through AI, particularly after witnessing the changes in productivity and production relations brought about by AI, the challenge became how to achieve meaningful large-scale commercialization of AI. This became their strategy for discovering new development models in the post-internet era. At this juncture, the key was to enable more consumer (C)-side users to comprehend and utilize AI as they do the internet.

During the Spring Festival, major players commenced promoting AI through traditional 'red envelope wars.' This proved efficacious in raising public awareness of AI. Now, more individuals are resorting to AI instead of traditional search methods to acquire content, products, and services. Subsequently, the shrimp-farming craze ignited by OpenClaw further propelled AI adoption, entering a new phase of development.

Whether through AI-themed red envelope wars or the nationwide shrimp-farming trend, it is evident that AI is supplanting the internet as a new lifestyle. This lifestyle shift is accompanied by an increased demand for tokens—ciyuan. With the impending Qingming Festival holiday, AI usage is becoming ingrained as a habit, leading to a surge in token calls. This phenomenon underscores the trend of AI becoming a new way of life.

II

As consumer demand for AI burgeons, we observe more players in the upstream AI industry accelerating industrialization and commercialization. They are leveraging AI to transform the traditional supply side, triggering an AI-driven supply-side reform.

AI supply is transitioning from laboratory stages to large-scale commercial deployment. At this stage, success hinges not on model power but on the ability to commercialize AI at scale. Only by achieving this can players meet the surging consumer demand for AI.

In this process, we witness reforms across various domains: Neolix's large-scale deployment of autonomous vehicles in Qingdao; Qiting's humanoid robot rentals securing new funding and recruiting urban operators; and Tencent Yuanbao's shift towards AI-powered social networking. These developments indicate that as consumer demand for AI products rises, the traditional internet-based traffic model is undergoing profound change. From an AI perspective, a new era of traffic dividends is emerging.

To capitalize on these new traffic dividends, players must expedite upstream AI supply through large-scale commercialization, thereby meeting consumer demand for AI. Consequently, demand for tokens—ciyuan—continues to escalate.

At this juncture, AI commercialization is no longer the exclusive purview of a few players but a necessity for all. The key to survival in the AI era lies in creating new products and services through AI. The increasing token calls directly reflect this trend.

III

At the China Development Forum 2026, Liu Liehong, Director of the National Data Administration, stated that propelled by technological innovation and commercial applications, the AI industry continues to expand. He also mentioned that the AI-related industry is anticipated to surpass RMB 10 trillion by the end of the 15th Five-Year Plan period. Behind this industrial scale lies the ongoing reshaping and reconstruction of industries by AI.

The sustained increase in token—ciyuan—calls directly mirrors this trend. As previously discussed, it reveals the growing demand for AI from both consumers and businesses, as well as the broader reshaping of industries, production, and daily life. Once complete, this transformation will usher in a new era of productivity and production relations built on AI.

According to reports from IDC and Gartner, by 2027, over 90% of enterprises will focus more on AI's impact, with 50% of large companies utilizing AI in at least one core business process. By 2028, 15% of daily work decisions will be made by AI.

Achieving this necessitates more than existing AI offerings. As demand for AI grows, its development is shifting from simple 'AI+' to 'AI×'—a phase where players must utilize AI to drive comprehensive, inside-out industrial transformation, not just meet basic daily needs.

For instance, in automotive manufacturing, Seres updates its car models annually, compared to the traditional 5-year minor and 10-year major update cycles. Its production lines can customize thousands of car variants, unlike traditional lines limited to 3-5 models.

In finance, AI applications have aided ICBC and CMB in constructing thousand-card AI training infrastructures, boosting customer service efficiency by 18% and tripling loan approval efficiency through AI risk control.

As AI is deployed in more scenarios, it is evolving from standalone applications to ecological collaboration. This shift still hinges on increased token—ciyuan—calls. As AI permeates various industries, we will witness even more new demands emerge.

Final Thoughts

The sustained increase in token—ciyuan—calls is merely a snapshot of China's AI industry undergoing profound transformation. It signifies AI replacing the internet as a new lifestyle, a new industrial force, and its development moving towards scale and ecological integration. This is just the beginning, as AI continues to unfold a new industrial landscape.