AT&T Business has extended the reach of its 400 Gbps (400G) wavelength service to more than a dozen new US metro locations, bolstering connectivity options for enterprise customers, including AI-focused operations.
The new network anchor points include Ashburn, Virginia; Atlanta; Chicago; Columbus, Ohio; Dallas; Denver and Englewood, Colorado; Kent, Washington; Los Angeles; Miami; New York; Portland, Oregon; San Jose, California; Seattle; Secaucus, New Jersey; and Tempe, Arizona. This pushes the total reach to more than 40 metro locations.
Those locations will now have access to AT&T’s wavelength edgeless handoff capabilities that link 440,000 “properties” serving more than 2.3 million business tenants. New connectivity services at up to 400G speeds can be instantiated in as little as 15 days.
AT&T tied the expansion to long-standing relationships with more than 130 carrier hotels nationwide, “leveraging an edgeless design that helps extend high-capacity wavelength connectivity without requiring traditional network Edge builds at every location.”
AT&T CFO Pascal Desroches recently noted that data center connectivity is reliant on the carrier’s network assets, with a particular reliance on AT&T’s fiber-based architecture to connect increasingly AI-heavy workloads.
“AI doesn’t exist without our connectivity … plain and simple,” Desroches said during the recent 2026 Mizuho Technology Conference. “And we know one thing for sure: data consumption is going to continue to go up under AI.”
Toward this, Desroches touted fiber as providing “the best ability to meet” that demand.
“You have symmetrical speed, highest speeds available, lowest latency, and importantly, the cost to deliver the unit cost to deliver bandwidth is the lowest with fiber,” Desroches said. “And it has a great maintenance profile, less power consumption, all things that make it incredibly attractive.”
This plays into AT&T’s fiber wheelhouse, which the carrier is currently building out toward passing 60 million premises by the end of this decade and driving an increase in enterprise-derived revenues.
“As we build out more fiber locations, and as we penetrate the existing fiber base, we’re seeing an acceleration in growth in business fiber revenues, both shared and dedicated,” Desroches said of that process.
The executive also took a dig at rivals that have more recently entered the fiber space.
“It’s a long game. You don’t wake up five years from now and say, ‘oh, I want to be in the fiber business,’” Desroches noted. “We’ve been at this for a while, and we are building a network that is AI-ready. There will be a lot ready to catch up and trying to figure out how to catch up as those consumption trends materialize. We’re not going to be in that position.”
March toward 400G
AT&T initiated its 400G work 10 years ago, starting with a trial that used open platform routers. That work quickly accelerated through specific vendor trials, including one with Juniper Networks in late 2019.
AT&T initiated the current 400G offering last year, touting the high-capacity fiber service as a way to accelerate deployment of data-intensive enterprise workloads across its nationwide wavelength network.
The fiber service runs on dense wavelength-division multiplexing (DWDM) technology, which has been able to transport up to 1.6 Tbps (1.6T) on a single wavelength in lab trials. That scalability positions AT&T’s backbone for the growing traffic generated by AI agents and distributed Edge applications.
Customers at that time could combine four 400G wavelength circuits using the new 1.6T coherent technology, though supporting equipment was sold separately and 400G services were available only in select markets.
Other telecom operators have jumped on the 400G push to help bolster connectivity capacity, including Lumen Technologies and Verizon.
More in Telecoms & 5G
