Verizon’s fiber network achieves new speed record in effort to enable AI apps

The race to prepare telecommunications infrastructure for the demands of artificial intelligence has reached a new milestone with Verizon’s successful high-speed fiber network trial in Boston.

Breaking new ground: Verizon has achieved speeds of 1.6 terabits per second (Tbps) in its live fiber network, doubling its previous record of 800 Gbps from 2020.

• The trial was conducted in Boston over a 118-kilometer metro fiber route with 10 hops
• The test utilized Ciena’s WaveLogic 6 Extreme coherent optical solution
• The network configuration included nine reconfigurable optical add-drop multiplexers (ROADMs), demonstrating performance in a complex network environment

AI-driven infrastructure demands: The upgrade represents a strategic move to prepare for the massive data requirements of artificial intelligence workloads.

• AI applications require real-time analysis of billions of data points
• Verizon has undergone a multi-year network transformation, including cloud-native architecture deployment and spectrum expansion
• The carrier aims to position itself as the preferred provider for AI workloads

Environmental benefits: The network upgrade offers significant sustainability advantages alongside performance improvements.

• Ciena’s WaveLogic 6 Extreme technology provides an 86% reduction in emissions per terabit of capacity
• The upgrade promises higher reliability alongside energy savings
• These improvements align with growing industry focus on sustainable infrastructure

Industry perspective: Network experts are anticipating dramatic increases in AI-related network traffic and infrastructure requirements.

• Monthly AI-enriched network traffic is projected to grow at a 120% compound annual rate through 2030
• A Ciena survey found 99% of telecom and IT engineers believe fiber network upgrades will be necessary to support AI traffic
• Similar 1.6 Tbps trials have been conducted with Telstra, Ericsson, and Arelion

Looking ahead: While the impact of AI on data center infrastructure is becoming clear, questions remain about broader network requirements.

• Uncertainty exists about which specific network segments will require capacity upgrades
• The distribution of increased data transmission needs across metro rings, long-haul links, and submarine cables remains unclear
• Successfully scaling AI will depend on secure, sustainable, and cost-effective data movement from core to edge data centers

Network evolution implications: The race to upgrade network infrastructure for AI workloads highlights both the opportunities and challenges facing telecommunications providers as they prepare for an AI-driven future, with questions remaining about optimal network architecture and capacity distribution across different segments of the network.