The competitive dynamics within the AI computing power sector are experiencing a significant transformation and reconfiguration. For the past two years, the industry has been ensnared in a GPU arms race, where companies amassing larger GPU inventories and constructing more extensive computing clusters have enjoyed a competitive edge. However, the prevailing paradigm of 'gaining dominance through sheer scale' is now showing signs of deterioration.

Under the Token-based billing model, the competitive rules have transitioned from 'hardware sales' to 'Token sales.' The incremental benefits derived from continuously adding hardware are gradually waning, and efficiency has emerged as the new centerpiece of competition. In early July, the Jing Suan Token Factory was inaugurated, and Trend Intelligence Technology secured cumulative financing surpassing 1 billion yuan within a mere six months. The former signifies the standardization of computing power provision, while the latter underscores capital's preference for optimization technologies. Collectively, these two milestones suggest that computing power optimization is transitioning from a supporting role to a leading position, becoming a decisive factor in corporate profitability and reshaping the domestic computing power landscape.

Nevertheless, the industry continues to grapple with challenges such as fragmented chip architectures, a scarcity of full-stack talent, and inadequate cluster stability. A standardized framework has yet to be established, and the protracted contest for computing power soft power has only just commenced.