The core of AI computing power is the GPU. With the rapid iteration of GPU products, it drives the acceleration of network connection solutions, and thus releases a long-term and continuous demand for the update and iteration of optical modules, which are key devices for optical interconnection. Currently, the data communication market has reached an inflection point. 800G and 1.6T will dominate the market in the next five years and will have a relatively long overlapping period.

Entering the 800G or even 1.6T era, as the number of internal channels in optical modules increases, the number of devices that need to be aligned and coupled increases proportionally with the number of channels. The industry chain must face the challenges brought by increased power consumption and costs. Multiple interconnection technology evolution paths, such as LPO, CPO, and IOI, have emerged one after another, and silicon photonics and heterogeneous integration have become the focus of attention. At the same time, due to the lag of standards, in the DCN scenario, the rapid deployment, reliability, and intelligent operation and maintenance of optical modules in AI intelligent computing centers have also received much attention. This webinar will conduct in-depth discussions on optical interconnection within intelligent computing centers, showcase the latest progress of optical interconnection technologies, introduce future development trends, and promote the innovative development of interconnection technologies for intelligent computing centers.