The drastic increase of bandwidth demands in AI datacenters requires new solutions with low power consumption and high bandwidth densities. Further increasing the total bandwidth with copper interconnects is challenging, thus it is crucial to introduce optics into the scale-up network. For this purpose, the most important metrics are the energy efficiency of the total link in addition to the spatial bandwidth-density and reach. With the inefficiency of copper traces at high speeds, transitioning to co-packaged optics is no longer optional but essential. Here, we show that the mature VCSEL technology offers the ideal combination of low-cost, low-latency, high-reliability, and energy efficiency at all bitrates, thanks to their unique versatility and high wall-plug-efficiency. This work has been submitted to the IEEE for possible publication. Copyright may be transferred without notice, after which this version may no longer be accessible.

翻译：人工智能数据中心对带宽需求的急剧增长，亟需兼具低功耗与高带宽密度的新型解决方案。通过铜互连进一步提升总带宽面临挑战，因此在扩展网络中引入光互连技术至关重要。为此，除空间带宽密度与传输距离外，链路整体能效成为最关键的性能指标。鉴于铜互连在高速传输下的效率瓶颈，向共封装光学技术转型已从可选项转变为必然选择。本文论证了成熟的VCSEL技术凭借其独特的通用性与高电光转换效率，能够在全比特率范围内实现低成本、低延迟、高可靠性与高能效的理想组合。本研究成果已提交IEEE审议出版，版权可能在不另行通知的情况下转移，此后当前版本或无法继续公开获取。