Video calling is technically one of the most demanding things the internet does routinely. That is not obvious from the user's perspective, because when it works it feels as simple as a phone call. But the infrastructure problem it solves is meaningfully harder than the one behind streaming, and understanding why reveals something specific about where data centers need to exist and why their location relative to the people using them matters more than most people realize.
Streaming and Video Calls Are Not the Same Problem
When you stream a movie, the system has an advantage: time. Your device can download several seconds of video ahead of what you are currently watching, buffering it locally so that small hiccups in the network stay invisible. If the connection stutters for half a second, you never notice, because your device was already ten seconds ahead.
A video call cannot do this. The person on the other end is speaking right now, and their voice and image need to reach you as close to instantly as the physics of data transmission allow. There is no buffer to absorb delay. If the signal takes too long to travel, the conversation becomes unworkable. People talk over each other because they cannot tell when the other person has finished speaking. Audio and video fall out of sync. The call deteriorates from a normal conversation into something that feels like communicating across a bad telephone line from several decades ago.
The technical term for this travel time is latency, measured in milliseconds. For most internet applications, a few hundred milliseconds of latency is entirely acceptable. For real-time voice and video, the threshold is much tighter. Once latency exceeds roughly 150 milliseconds in each direction, the natural rhythm of human conversation begins to break down noticeably.
Why Distance Is the Problem
Data travels through fiber optic cables at approximately two-thirds the speed of light. That sounds fast, and over short distances it is. But the speed of light across a continental landmass is not instantaneous. A signal traveling from a user in Phoenix to a server in New York and back covers roughly 4,500 miles round trip. At two-thirds the speed of light, that physical journey alone accounts for around 40 milliseconds of latency, before accounting for the additional time added by routers, switches, and network processing along the way.
For a video call connecting two people in the same metropolitan area, the round trip is far shorter and latency stays comfortably within the range where conversation feels natural. For a call connecting someone in suburban Atlanta to a colleague in another city, the data is likely routing through one or more intermediate facilities. The location of those facilities, and how efficiently the network between them operates, directly determines whether the call feels like a normal conversation or a frustrating exercise in talking through delays.
This is why the phrase "the internet" is slightly misleading. There is no single network. What exists is a web of interconnected infrastructure, with data routing through whichever available path reaches the destination fastest. The quality of any real-time communication depends heavily on which physical facilities that data passes through and how far it has to travel to get there.
What Edge Infrastructure Actually Does
The response the industry developed to the latency problem is called edge computing, and its logic is straightforward. If the problem is that data has to travel too far, the solution is to move processing and routing closer to the people generating and receiving it. Rather than routing a video call from a suburb of Charlotte through a facility in Northern Virginia or Chicago, edge infrastructure places smaller, strategically located facilities inside or near major metropolitan areas, shortening the physical path the data has to travel.
These facilities are not always the massive campuses that most people picture when they hear "data center." Many are smaller buildings, purpose-built to handle regional traffic routing and processing, positioned based on where large concentrations of users are located. Their value is not size. Their value is location.
The Building Your Call Passed Through
The next time a video call holds together cleanly through ninety minutes of screen sharing and sidebar conversation, that outcome was made possible by a specific set of buildings in specific places. One of them may have been several states away, handling long-haul routing between major network hubs. One was almost certainly closer, somewhere in the metro region, handling the regional distribution of data before it reached your building and your device.
The buildings supporting this are typically unremarkable structures near utility corridors and fiber routes, drawing power steadily and operating continuously. The region they sit in has access to reliable, low-latency connectivity partly because they exist there. The region that does not have them relies on infrastructure further away, and the difference is measurable in milliseconds that accumulate into noticeable degradation in communication tools millions of people now depend on for work every day.
This is the third article in The Daily Connection, a series by Blueprint Data Centers on the physical infrastructure behind everyday digital life. Blueprint is an independent data center platform developing greenfield data centers designed with flexibility to support a range of use cases including high-performance computing, AI and other advanced workloads. Follow the series for plain-language explanations of the infrastructure communities use every day.
