Verizon's AI Gold Rush Should Not Come at the Consumer's Expense

Verizon is currently chasing a new kind of gold: the explosive compute demands of Silicon Valley. During a second-quarter 2026 earnings call, CEO Dan Schulman unveiled "AI Connect," a business initiative designed to monetize the company's infrastructure through dark fiber deals and the conversion of central offices into mini-data centers.

As Ars Technica first reported, the centerpiece of this strategy is a deal worth more than $1 billion to link Google data centers using Verizon's dark fiber routes. Schulman isn't stopping there; he indicated that further AI-related deals, potentially worth multiple billions of dollars over several years, could be announced by the end of the year.

From a corporate balance sheet perspective, this looks like a masterstroke. By selling unused dark fiber capacity to hyperscalers, Verizon can recoup some of the costs associated with its $20 billion acquisition of Frontier Communications, which expanded its fiber footprint from 15 states to 31 states and Washington, DC, at the start of 2026. Furthermore, the company is retrofitting central offices—the very hubs that house the physical equipment for our mobile and internet networks—into small data centers to support "inference edge computing." Schulman claims a small trial of this capability sold out in just 24 hours, catering to high-tech needs like autonomous robotaxi fleets and robotics.

But as a consumer advocate, I find this pivot concerning. While Schulman speaks of optimizing the "ever-exploding need for compute" and enabling ultra-low latency for remote surgery and autonomous driving, we must ask: who is this network actually for?

Verizon is aggressively pivoting toward corporate AI workloads while simultaneously slashing its human footprint. Ars Technica reports that after laying off 13,000 workers at the end of 2025, the company recently transferred 2,500 employees and 274 retail stores to franchise owners, while cutting another 500 corporate positions.

There is a dangerous tension here. When a carrier prioritizes the "inference edge" for robotaxis and the data-center connectivity of Google, the risk is that the core mobility and broadband services—the ones regular people pay for every month—become secondary. If the physical equipment in central offices is being repurposed to serve AI workloads, we need guarantees that the reliability of the consumer network isn't being compromised to make room for the next AI trend.

Verizon expects AI Connect to substantially grow revenues and margins over the next five to 10 years. However, corporate growth should not be subsidized by a degradation of service or a lack of investment in the consumer experience. We cannot allow the pursuit of AI profits to turn the essential connectivity of the American public into a secondary priority. The "gold rush" is fine for the shareholders, but the network must remain reliable for the people who actually pay the bills.