The internet handles roughly 120 exabytes of data per month, though the exact number shifts constantly as more people stream video and use cloud services

An exabyte is one billion gigabytes. To put that in perspective: if you downloaded one gigabyte per second without stopping, it would take you 31 years to read one exabyte. The internet moves that amount roughly 120 times every month.

This number matters because it shows why your internet connection sometimes slows down during peak hours, why data centers need enormous amounts of electricity, and why internet providers keep upgrading their infrastructure. The data moving through the internet is not evenly distributed — most of it is video streaming, cloud backups, and social media, not email or web browsing.

The total keeps growing. Five years ago the figure was closer to 60 exabytes per month. In another five years it will likely be double what it is now. But the number itself is less useful than understanding what kinds of data make up that total and where the bottlenecks actually happen.

Key Takeaways

  • Video streaming accounts for roughly half of all internet traffic, with Netflix, YouTube, and TikTok consuming far more bandwidth than any other services.
  • Data does not travel as one continuous stream — it breaks into small packets, each tagged with where it came from and where it is going, then reassembles at your device.
  • The internet has no single "size" because data is constantly being created, deleted, and moved between devices, so measurements capture only what is flowing at one moment.
  • Most internet congestion happens at the "last mile" — the connection from your internet provider's equipment to your home — not on the backbone cables that carry data between cities.

Where the data actually comes from

The 120-exabyte figure counts data flowing between devices, not data stored somewhere. Every time you watch a video, send an email, upload a photo, or load a webpage, that counts. Every time a security camera sends footage to the cloud, every time your phone backs up to iCloud or Google Drive, every time a company's servers talk to each other — all of it adds up.

Video is by far the largest piece. Streaming services, social media videos, video calls, and live broadcasts together make up roughly 50 to 60 percent of all internet traffic. A single Netflix stream uses between 1 and 7 gigabytes per hour depending on quality. If millions of people are streaming simultaneously during evening hours, that alone can strain the entire system.

The rest breaks down roughly like this: cloud services and backups (around 15 percent), social media and web browsing (around 15 percent), gaming and other applications (around 10 percent), and everything else including email, messaging, and business traffic (around 10 percent). These percentages shift by region and time of day — business districts see more work-related traffic during office hours, while residential areas peak in the evening when people come home and stream.

How data actually moves through the internet

Data does not travel as one continuous stream from your device to its destination. Instead, your device breaks it into small chunks called packets, each typically between 1,500 and 65,000 bytes. Each packet gets a header that contains the sender's address, the destination address, and instructions for reassembly.

These packets do not all take the same route. When you request a webpage, one packet might travel through cables in your city, another through a fiber line across the country, and a third through an undersea cable to Europe and back. Routers along the way read the destination address and forward each packet toward its target. If a cable is congested, a router sends packets down a different path. This is why the internet keeps working even when parts of it fail — there are usually multiple routes available.

At the destination, your device reassembles the packets in the correct order and reconstructs the original data. This happens so quickly that you do not notice the packets are separate — a webpage loads as one continuous image, a video plays smoothly, a message appears when ready. But underneath, thousands of packets are arriving out of order and being reassembled constantly.

Why the internet does not have a fixed size

You cannot measure the internet the way you measure a hard drive. A hard drive has a fixed capacity — 256 gigabytes, one terabyte, whatever. The internet has no fixed size because data is constantly being created, deleted, and moved.

When you delete an email, that data is gone. When you upload a photo to Instagram, that data now exists on Instagram's servers. When you stream a movie, Netflix sends you the data, but the movie still exists on Netflix's servers for the next person to watch. The 120-exabyte figure measures only the data in motion at any given moment, not the total amount of data that exists somewhere on the internet.

This is why measurements vary depending on who is counting and when. A measurement taken at 3 a.m. on a Tuesday will be much lower than one taken at 8 p.m. on a Friday. Measurements also depend on what you count — some include only public internet traffic, others include private corporate networks and cloud services, which adds significantly to the total.

Where congestion actually happens

Most people assume the internet slows down because the main cables are full. In reality, the backbone — the high-capacity fiber optic cables that carry data between cities and countries — has plenty of capacity. The bottleneck is usually much closer to home.

The last mile is the connection from your internet provider's equipment to your house. If you have cable internet, that last mile is shared with your neighbors. If 50 households on your street all stream video at the same time, the shared cable gets congested and everyone's speed drops. Fiber internet has more capacity per household, which is why it is faster and more reliable. Satellite internet has high latency (delay) because data has to travel to space and back, not because there is not enough total capacity.

During peak hours — typically 7 p.m. to 11 p.m. — internet providers see their heaviest traffic. This is when most people are home and streaming. Providers have to balance the cost of upgrading infrastructure against the fact that most of the day that infrastructure sits underused. This is why your internet is usually faster at 2 a.m. than at 8 p.m., even though the backbone itself is not congested.

How data centers handle the volume

The companies that move all this data — Google, Amazon, Netflix, Facebook, and others — operate massive data centers in dozens of locations. These are not small server rooms. A large data center can occupy several hundred thousand square feet and house hundreds of thousands of servers.

These facilities use enormous amounts of electricity to power the servers and cool them. Data centers account for roughly 1 to 2 percent of global electricity use, and that percentage is growing. Companies are constantly building new data centers closer to population centers to reduce the distance data has to travel, which reduces latency and lets them serve content faster.

The servers in these data centers are not like the computer on your desk. They are specialized machines designed to handle thousands of requests simultaneously. A single server might handle requests from millions of users per day, storing pieces of data, retrieving it, and sending it back out across the internet.

Why the numbers keep growing

Internet traffic has roughly doubled every two to three years for the past two decades. This is not because people are using the internet more hours per day — it is because the kinds of things people do online have changed. Ten years ago, most internet use was text and images. Now it is mostly video.

A single high-definition video stream uses thousands of times more data than a text email. As video quality improves (4K, 8K), as more devices connect to the internet (smart home devices, security cameras, cars), and as cloud services become the default way people store files, the total keeps climbing.

This growth is why internet providers keep upgrading their networks and why companies keep building new data centers. The infrastructure has to expand to handle the demand, or the internet slows down for everyone.

Frequently Asked Questions

Is there a limit to how much data the internet can handle?

Not a hard limit, but a practical one. The internet can handle whatever capacity the infrastructure supports. When demand exceeds capacity in a particular area, that area gets congested. Providers upgrade the infrastructure to increase capacity, which costs money and takes time. So the real limit is how much providers are willing to invest, not a technical maximum.

Does my internet provider know how much data I use?

Yes. Your provider can see how much data flows to and from your connection. Some providers charge based on data usage (metered billing), while others offer unlimited plans. Even with unlimited plans, your provider tracks your usage and may slow your connection if you exceed a threshold during peak hours.

Why does video use so much more data than other things?

A video is thousands of still images shown in rapid sequence. Each image contains millions of pixels, and each pixel contains color information. Even compressed video requires far more data per second than text or audio. A two-hour movie in high definition can be 5 to 10 gigabytes, while a two-hour audiobook is typically 100 to 200 megabytes.

Does the internet ever run out of bandwidth?

Specific connections can run out of bandwidth — your home internet, a particular cable, a data center's connection to the backbone. But the internet as a whole does not run out because new infrastructure is constantly being built. When demand exceeds capacity somewhere, that becomes a business problem for the provider, not a technical problem for the internet itself.

How much of the internet is stored on servers versus traveling at any moment?

The vast majority is stored. The 120-exabyte figure measures data in motion. The total amount of data stored on servers, hard drives, and cloud services worldwide is measured in zettabytes — thousands of exabytes. Only a tiny fraction of that is actively traveling at any given moment.