Search
As the application of AI agents continues to deepen, the volume of Token calls has skyrocketed, propelling the reduction of computing power costs to the forefront of attention for all stakeholders. The industry is now refining the utilization of large models via computing power platforms and expediting the construction of large-scale computing clusters. Among these initiatives, ultra-large-scale computing clusters underpinned by domestic computing power chips are widely regarded as the linchpin for cost reduction. It is anticipated that within the next three to five years, cutting-edge technologies such as optoelectronic integrated chips will be deployed. These chips boast lower computational latency and power consumption, and are projected to slash the cost per unit of Token by 50% or even more.
-
C114 Communication Network -
Communication Home
