Verizon said a Verizon Google dark fiber agreement worth more than $1 billion will connect Google data centers, giving the phone company a fresh way to sell network capacity into the AI buildout. CEO Dan Schulman disclosed the deal during Verizon’s second-quarter 2026 earnings call, along with a broader “AI Connect” push aimed at data-center links and low-latency edge computing.
The Google contract is the first named win in what Verizon is pitching as a larger AI infrastructure business. Schulman told investors that Verizon expects to announce more AI-related agreements by the end of the year, with those deals worth multiple billions of dollars in revenue over the next several years.
Verizon’s play is not mysterious: hyperscale AI systems need fast, high-capacity links between data centers. Dark fiber is fiber-optic cable that has been installed but is not actively carrying traffic. A customer can light it with networking equipment and use it as dedicated capacity between sites, rather than buying a conventional managed service.
What is Verizon selling to Google?
Verizon said Google will use Verizon dark fiber routes to connect data centers. The company did not disclose the specific routes, data-center locations, term length, or technical capacity covered by the more-than-$1 billion deal.
Schulman framed the demand as a shift from merely serving consumer broadband and mobile users to providing the connective tissue between large compute facilities. That is convenient framing, but the mechanism is real enough: AI training and inference can require moving large volumes of data among specialized facilities, and private fiber routes are one way to reduce bottlenecks between them.
Verizon is also trying to turn some of its own central offices into small data centers. Central offices are the buildings that house equipment for internet and mobile networks. Schulman said Verizon is in the early stages of converting some of those sites into “remote data centers” for inference workloads that need low latency, including robotics, remote surgery and autonomous driving applications.
According to Schulman, Verizon ran a small trial of that edge-computing capacity and sold it out within 24 hours. The company did not provide the trial’s size, customer names, pricing, or workload details, so the claim is best read as an early sales signal rather than proof of a mature business.
The central-office plan overlaps with work Verizon was already doing. The company has been removing copper wiring from those facilities and replacing legacy copper network assets with fiber-optic technology for broadband services. Fiber can carry more data and use power more efficiently than old copper lines, which makes the cleanup useful even without an AI label slapped on the rack.
Verizon also has more fiber to sell after closing its $20 billion acquisition of Frontier Communications at the start of 2026. Verizon said that deal expanded its fiber footprint from 15 states to 31 states plus Washington, DC. Selling unused fiber capacity to Google and other AI companies could help monetize that enlarged network.
The AI push comes while Verizon is still cutting costs elsewhere. The company laid off 13,000 workers at the end of 2025, leaving it with just under 90,000 employees at year-end, according to Fast Company. More recently, Verizon transferred 274 retail stores and 2,500 employees to franchise owners and cut another 500 corporate jobs, Fast Company reported.
Schulman told investors that AI Connect deals should start affecting Verizon’s revenue and margins next year and grow over the next five to 10 years. For now, the confirmed business is narrower: one large Google dark fiber deal, some early central-office conversions, and a carrier trying to make old network real estate useful to the AI boom.
This story draws on original reporting from Ars Technica.
