Verizon lifts annual forecasts, signs more than $2 billion AI infrastructure deal with AWS

Verizon made two major announcements. First, it raised its forecast for 2025 postpaid phone additions to between 1.1 million and 1.2 million. It also increased its adjusted earnings-per-share forecast to $5.04-$5.06. These upgrades reflect confidence in demand for its simpler wireless plans.

Second, Verizon signed an agreement worth more than $2 billion with Amazon Web Services. Verizon will provide high-capacity fiber-optic connections for AWS data centers. Those links are designed to move data faster and more reliably as AI workloads expand.

The announcements show Verizon pursuing two growth paths. Its wireless business is seeking more subscribers through transparent pricing and unlimited data. Its network business is targeting large technology customers. Verizon also announced a separate, more than $1 billion dark-fiber agreement with Google in July, suggesting AI infrastructure could become a larger revenue source.

Postpaid phone subscribers are customers who use a wireless plan and receive a bill afterward, usually each month. They differ from prepaid customers, who pay before using service. For a carrier, postpaid additions measure growth in its recurring customer base. Strong additions can signal that pricing, network service, and promotions are attracting customers.

Verizon reported 406,000 monthly-bill-paying wireless additions in the third quarter. That was nearly double Visible Alpha’s estimate of 209,439. Verizon linked its stronger performance to demand for simplified plans, including its Simplicity plan with transparent pricing and unlimited data.

This measure matters because customer growth supports future service revenue. Verizon raised its expected annual postpaid phone additions to 1.1 million-1.2 million, from 750,000-1 million. However, subscriber growth does not guarantee higher total revenue, as weaker device upgrades later showed.

Verizon raised its expected annual postpaid phone additions from 750,000-1 million to 1.1 million-1.2 million. That lifts the forecast’s lower end by 350,000 and its upper end by 200,000. The change reflects stronger-than-expected wireless subscriber gains and confidence in simplified plans.

The company also raised its full-year adjusted earnings-per-share forecast. The previous range was $4.99-$5.04. The new range is $5.04-$5.06. Verizon therefore increased both ends of its earnings outlook, though the ranges overlap at $5.04.

Separately, Verizon signed an AI infrastructure agreement worth more than $2 billion with Amazon Web Services. Verizon will provide high-capacity fiber-optic connectivity for AWS data centers. The network will support faster, more reliable transfers for growing AI computing workloads, adding a potential business channel beyond traditional wireless and broadband services.

Subscriber growth and revenue can move differently. Verizon added many more monthly-bill-paying wireless customers than analysts expected, but its total third-quarter revenue was $34.3 billion. That fell below the $34.74 billion analyst estimate. The shortfall came mainly from lower device upgrade volumes.

Customers held on to their phones longer after Verizon began cutting device subsidies. Fewer upgrades mean fewer device sales and related revenue, even when more people are paying monthly wireless bills. In this case, the stronger customer count did not fully offset weaker equipment activity.

The result highlights a trade-off in Verizon’s strategy. Simplified offerings helped attract subscribers, while reduced subsidies changed customer upgrade behavior. Verizon still raised its annual subscriber and earnings forecasts. However, the quarterly revenue miss shows why subscriber growth alone does not describe the whole financial picture.

Verizon will provide high-capacity fiber-optic connectivity for Amazon Web Services data centers. Fiber links can carry large amounts of information quickly and reliably. That gives AWS stronger connections for moving data into, out of, and between facilities supporting AI computing workloads.

The key mechanism is network capacity and dependability. AI workloads can involve large volumes of data moving between storage and computing resources. Verizon’s agreement is intended to enable faster, more reliable data transfers. The article does not provide technical performance figures, but it directly links the connectivity to AWS’s growing AI needs.

For Verizon, the deal creates a business opportunity beyond standard wireless and broadband services. It follows a more than $1 billion agreement with Google to connect data centers using dark fiber. Together, the agreements show how rising AI demand could create new revenue streams for telecom operators.

Dark fiber generally means fiber-optic cable that has been installed but is not currently carrying active communications signals. A customer can use dedicated strands by adding the necessary transmission equipment. This arrangement can provide control over capacity and network design, although the article does not describe Verizon’s technical setup.

Verizon announced a more than $1 billion agreement with Google to connect data centers using dark fiber. The connection is part of the telecom company’s effort to serve infrastructure needs created by expanding AI activity. The AWS agreement uses high-capacity fiber-optic connectivity for a similar data-center purpose.

Telecom companies already operate extensive fiber networks. Selling connectivity to data centers can let them earn revenue from those networks beyond traditional wireless and broadband services. Verizon’s agreements with Google and AWS show why AI infrastructure may become a new commercial opportunity, although the article gives no forecast for its eventual size.

Modern AI systems generally need large data centers because their workloads require substantial computing and storage resources. Training and running these systems involves processing large datasets and coordinating many computing machines. Concentrating that equipment in data centers provides the scale needed for intensive AI workloads.

The network is the connection between those resources. Data must move between storage and computing systems, and sometimes among separate data centers. High-capacity fiber can carry more data, while reliable connections reduce disruptions. Verizon’s AWS agreement is intended to support faster, more reliable transfers for growing AI computing workloads.

This infrastructure helps explain why telecom companies see AI as a possible new business area. Verizon will connect AWS data centers and separately agreed to connect Google data centers using dark fiber. The article does not quantify AI’s future network requirements, but it identifies rising AI demand as a potential revenue opportunity.