Cisco and Nvidia are teaming up to address a critical challenge in the AI industry: the growing data center bottleneck. The collaboration focuses on developing next-generation optical networking technology capable of managing the massive surge in data traffic driven by AI applications. Cisco President and Chief Product Officer Jeetu Patel highlights that AI agents, due to their complexity, could demand significantly more network bandwidth than human-led processes performing the same tasks.
The new technology aims to enhance the efficiency and speed of data transfers within AI data centers, allowing hundreds of thousands of GPUs to operate in unison. This innovation is expected to reduce the operational costs associated with AI, making the technology more accessible and scalable. The partnership underscores the urgency to upgrade data center infrastructure to keep pace with the rapid advancements in AI.