Verizon touts $1B dark fiber deal for Google data centers as first of many

2026-07-27 · Source: AI - Ars Technica · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Cloud Computing & IT Infrastructure, Robotics & Autonomous Systems · Depth: Fundamental Awareness, quick

Summary

Verizon launched its "AI Connect" business initiative, securing a deal worth over \$1 billion to connect Google data centers using its dark fiber routes. This move, announced during the Q2 2026 earnings call, is part of a broader strategy to capitalize on AI spending, with CEO Dan Schulman anticipating additional multi-billion dollar AI-related deals by year-end. Concurrently, Verizon is converting some central offices, previously housing copper wiring, into small data centers to support ultra-low latency AI inference workloads at the edge, targeting applications like robotics and autonomous driving. This expansion follows Verizon's \$20 billion acquisition of Frontier Communications at the start of 2026, which significantly grew its fiber footprint from 15 to 31 states, and comes amidst ongoing job cuts. The company expects AI Connect to impact revenues and margins starting next year, growing substantially over the next five to ten years.

Key takeaway

For Directors of AI/ML or VPs of Engineering evaluating infrastructure for distributed AI, Verizon's "AI Connect" initiative signals a significant shift. You should consider telecom providers as strategic partners for both hyperscale data center interconnects and low-latency edge inference. This trend suggests that leveraging existing fiber networks and purpose-built edge data centers will be crucial for deploying demanding AI applications like autonomous systems or remote robotics efficiently.

Key insights

Verizon is strategically repurposing its extensive fiber infrastructure and central offices to become a key enabler for hyperscale and edge AI computing.

Principles

• Optimize data center connectivity for exploding compute needs.
• Edge AI inference demands ultra-low latency infrastructure.
• Existing telecom assets can be repurposed for AI infrastructure.

Method

Verizon's method involves selling dark fiber capacity to hyperscalers and retrofitting central offices by removing copper, upgrading to fiber, and converting them into small data centers for ultra-low latency edge AI inference.

In practice

• Deploy dark fiber for hyperscale data center interconnects.
• Establish remote data centers for low-latency edge AI.
• Upgrade legacy copper networks to fiber for AI readiness.

Topics

• Dark Fiber
• Edge AI
• Data Center Interconnects
• AI Inference
• Telecom Infrastructure
• Network Modernization

Best for: CTO, Executive, AI Architect, Director of AI/ML, VP of Engineering/Data, Investor

Related on AIssential

• Verizon CEO Eyes Billions in AI Infrastructure Opportunity - Broadband Breakfast — AI infrastructure demand offers a "once in a lifetime" revenue opportunity for network providers.
• Laser Chip Brings Multiplexing to AI Data Centers — Integrating DWDM lasers directly onto silicon wafers revolutionizes AI data center networking by enhancing bandwidth and reducing latency.
• Microsoft accelerates telecom return on intelligence with a unified, trusted AI platform — Telecoms achieve significant ROI from AI by integrating a unified, trusted platform across operations, networks, and customer experiences.
• NVIDIA GTC Telecom Special Address: The AI Grid—Intelligently Connecting AI Infrastructure — The AI grid merges telecom and AI stacks, creating a generational opportunity for AI-native infrastructure and services.
• Connectivity Revolution or Evolution Inside Data Centers? — AI workloads necessitate a revolutionary shift to optical, scale-out intra-data center networks, integrating optics closer to compute.
• How Google Is Quietly Financing the AI Data Center Boom — Hyperscalers use financial guarantees to de-risk AI infrastructure investments, accelerating market growth and securing their supply chains.

See Counsel's argued verdicts on the open AI decisions leaders are weighing →

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by AI - Ars Technica.