Partnership Targets Low-Latency AI at Internet Exchange Points
January 15, 2026
Source: broadbandbreakfast.com
Hunter Newby, co-CEO of IXP.us, was quoted in this Broadband Breakfast article on the QAI Moon partnership deploying low-latency AI compute at Internet Exchange Points.
by Sergio Flores
Moonshot and QumulusAI announced a strategic agreement with Connected Nation Internet Exchange Points, or IXP.us, to deploy a nationally distributed artificial intelligence compute and internet exchange platform designed to support low-latency AI workloads at the network edge.
Under the agreement, the companies will pair carrier-neutral internet exchange points with modular "AI Pods," initially targeting deployment at 25 IXP.us sites, with plans to scale to as many as 125 locations across U.S. research university campuses and municipalities over the next five years.
"With this strategic relationship, we will enable the first scalable low-latency compute infrastructure directly adjacent to our network-dense interconnection facilities, including on many research university campuses," said Hunter Newby, co-CEO of IXP.us.
The platform is intended to address growing demand for latency-sensitive AI applications by colocating GPU compute closer to networks, data sources, and end users, offering an alternative to centralized hyperscale data centers.
"This partnership represents the physical convergence of power, compute, and interconnection at the exact point where AI demand is moving," said Ethan Ellenberg, CEO of Moonshot. "By pairing modular infrastructure with carrier-neutral interconnection, we're creating a repeatable architecture for deploying AI closer to the network edge."
The first deployment is scheduled to begin by July 2026 at an IXP.us site on the Wichita State University campus in Kansas, with additional markets planned as sites come online.
"AI workloads are increasingly inference-driven and latency-sensitive, but the infrastructure hasn't kept pace," said Mike Maniscalco, CEO of QumulusAI. "This partnership allows us to place GPU compute directly at the network edge, where data moves and decisions happen."
Executives said the model is designed to support network operators, enterprises, and AI developers seeking lower latency, greater infrastructure flexibility, and broader access to AI compute instead of relying on large, centralized data centers.
