AI integration reshapes network infrastructure and operations

The integration of artificial intelligence into network infrastructure and operations is accelerating. At the APNIC 62 conference, discussions focused on how AI-driven tools, such as agentic AI and machine learning, are reshaping network design, data center architecture, and autonomous management. Experts highlighted the shift toward next-generation optical fabrics and predictive fault detection to handle increasingly complex AI workloads.

Industry developments show a growing reliance on agentic AI, with a Cisco survey indicating that 51% of organizations are already running such technology in production. Companies like Selector and NetBrain are introducing AI agents to enhance network observability and problem identification. Additionally, Google is exploring the feasibility of orbiting AI data centers by planning to launch Tensor Processing Units (TPUs) into space.

Security remains a critical concern as traditional firewalls may lack the semantic intelligence required to secure modern AI traffic. Meanwhile, geopolitical and supply chain considerations continue to influence the sector, including Chinese government reviews regarding dependence on Broadcom networking silicon.