A survey of U.S. cities focusing on Chicago and Roswell, New Mexico, highlights nuances in cloud responsiveness as latency becomes an important factor in AI performance. Network latency is part of the path necessary to deliver AI computations from hyperscaler data centers to end users, and time-sensitive AI applications will need low latency connections. Cloud latency can be impacted by geographic location, internet routing infrastructure, and operators’ policies. The relatively narrow 42 ms gap between the lowest measured median cloud latency (104 ms in Providence) and the highest (146 ms in Roswell) highlights the depth of the internet’s fiber backbone. Oracle’s endpoints typically register higher latency measurements than AWS, Google Cloud Platform, and Microsoft’s Azure. Chicago, a major internet hub, notched median latency measurements that were in some cases 40 ms or lower than Roswell in the second quarter of 2026. However, Roswell enjoys a geographical routing advantage for westward traffic. Verizon’s 5G users in Roswell generally experienced snappier West Coast median cloud latency measurements in the second quarter of 2026, while AT&T and T-Mobile showed some faster eastbound connections. In Chicago, T-Mobile often notched the fastest median cloud latency across both eastern and western destinations.