How AI is transforming network management and infrastructure

Artificial intelligence (AI) is becoming indispensable in network management, offering operational efficiencies, increased reliability, and enhanced user experiences. From small businesses to global telecom giants like Swisscom, AI is being leveraged to reduce downtime, cut costs, and improve service quality. Swisscom’s partnership with Cisco-owned Outshift highlights the potential of agentic AI to redefine network operations.

The Nvidia Perspective

Nvidia CEO Jensen Huang has dubbed AI the "single most consequential technology in history," emphasizing its role in transforming corporate intelligence into digital intelligence. Enterprises are now digitizing their domain-specific intelligence, creating a perpetual AI lifecycle that enhances productivity and scalability.

AI for Networking vs. Networking for AI

• AI for Networking: Focuses on empowering IT admins to handle escalating workloads, including cybersecurity threats and new app deployments.
• Networking for AI: Involves building robust infrastructure to support AI workloads, such as GPU clusters and high-bandwidth networks. HPE Aruba Networking’s David Hughes notes the need for scalable, agile networks to handle AI’s dynamic demands.

Challenges and Solutions

• Network Overload: AI-generated content, like video and text, strains networks. BT’s CTO Colin Bannon warns of "elephant flows"—large data transmissions that require high bandwidth and robust congestion management.
• Edge Computing: NTT Data’s Tsuzumi, a small language model, addresses privacy and sustainability concerns by processing data locally at the edge.

Roadblocks to Adoption

Despite enthusiasm, 88% of UK business leaders face hurdles like poor infrastructure and unrealistic expectations. Expereo’s study reveals that 49% of UK networks struggle to support AI projects, underscoring the need for strategic investments in connectivity.

Key Takeaways

• AI is critical for modern network management but requires tailored infrastructure.
• Small language models like Tsuzumi offer edge solutions to reduce network strain.
• Businesses must balance AI’s potential with realistic goals and robust networks.

For more details, explore Nvidia’s vision or .