According to him, the so-called “war of a hundred models,” which shaped the sector’s development for several years, is gradually coming to an end. The market is consolidating around a limited number of major players capable of turning consumer AI products into reliable enterprise infrastructure.

Utility comes to the forefront

Yao noted that Chinese models may still slightly lag behind leading U.S. counterparts in some individual metrics, but this is becoming less decisive for commercial success within the country. What matters much more is the technology’s ability to solve real business problems.

According to the analyst, models have already reached a level where they can perform a significant share of intellectual work. Now the key question is not further performance growth, but the creation of products that users and companies are willing to pay for.

Chinese companies test new monetization models

Major technology platforms have already begun moving toward paid services. As an example, Yao cited ByteDance’s AI app Doubao, which launched several subscription plans in May priced from 68 to 500 yuan per month, or about 10 to 74 dollars.

The expert believes concerns about Chinese users’ unwillingness to pay for software are exaggerated. According to him, demand will primarily depend on whether services can demonstrate tangible value to customers.

The enterprise market will determine the winners

JPMorgan expects the next phase of China’s AI development to be linked to enterprise products, process automation, and the integration of AI into existing digital ecosystems. The bank’s analysts believe the industry is moving from competition over model quality to a battle for real revenue sources and long-term competitive advantages.

As the market consolidates, companies that can not only create advanced models but also build sustainable business models around them, focused on enterprise clients and practical technology applications, will gain the advantage.