Verizon: A Case Study in Telco Automation, AI and AN - How Does One of the World's Biggest Telcos Cut 33% of its Workforce in Six Years and Maintain Market-leading Margins?

Verizon has shed 33% of its workforce since 2019 while scaling to 1,000+ AI models in production and running one of the most advanced automation programs in telecom. While the public focus is on cost efficiencies and headcount, Verizon also aims to monetize AI through its AI Connect program. This profile documents what Verizon has built, what we can learn from its trials, and where more progress is needed.

Verizon reported revenue of $138.2 billion (B) in 2025, up 2.5% YoY. Its EBIT margin was 21.1%, above the global telco average of 15-16%. Just after the end of the year, Verizon closed its Frontier Communications acquisition, the biggest wireline deal in the US in a decade. Verizon targets $1B in annual opex synergies by 2028. Total headcount fell from approximately 135,000 in 2019 to 89,900 at year-end 2025. That’s a 33% reduction over six years, and the company announced a further 13,000-plus position cut in 4Q25. Labor cost per employee rose to ~$160K per year in 2023–2024 before moderating to roughly $140K in 3Q25, reflecting a workforce shift toward lower-cost, non-union staff. Labor cost as a share of opex (excluding depreciation) fell from 19.5% in 2019 to 15.9% at 3Q25, roughly six percentage points below the global average of 21.9%. This gap has widened every year. Verizon’s capex intensity was 12.6%, well below global averages, and will dip further to 11.6-12.0% in 2026 even as the Frontier fiber buildout accelerates.

Verizon is a global leader in TAIA, scoring 4.0 out of 5 in our review. It has depth in operational AI deployment and Open RAN. Its 1,000+ AI models in production, commercial multi-vendor RIC, formal AI Council and governance registry, and Running on ODA accreditation for the North Star Architecture add up to an impressive program. Where the company falls short: headline performance figures are self-reported without independent verification, no autonomous networks level target or ANLAV certification has been stated, and agentic AI remains in Catalyst consortium research rather than production deployment. For vendors, Verizon’s multi-vendor procurement posture across RAN (Samsung, Ericsson, Qualcomm) and AI infrastructure (Google Cloud, AWS, NVIDIA, Vultr) signals a deliberate avoidance of single-vendor dependency; competitive positioning based on standards compliance, interoperability, simplicity, and price are the more decisive factors. The AI Connect product suite and the Frontier wireline integration represent the two near-term areas where new vendor engagements are most likely to develop.

Organizations mentioned

• Alphabet (Google Cloud)
• Amazon Web Services (AWS)
• Amdocs
• AOL
• Apollo Global Management
• AT&T
• BSNL
• China Mobile
• Chunghwa Telecom
• Ericsson
• FCC
• Frontier Communications
• Intel
• Linux Foundation
• Meta
• NTT
• NVIDIA
• Oi
• Orange
• Qualcomm
• Red Hat
• Samsung
• Telecom Italia
• TM Forum
• T-Mobile
• Vultr
• Yahoo!