Does a VPN slow down your internet? Measure it yourself

Yes, a little. A VPN adds an extra stop to every connection, a few dozen bytes to every packet, and some encryption work to your device, so some loss is normal. How much depends on your line, your device and the server's distance. No single percentage applies to everyone, so the useful answer is to measure your own.

This post gives you the four causes, a ten-minute measurement you can trust, and what to change if the number is bad.

Why does a VPN slow down your internet?

1. The detour. Without a VPN, your traffic goes from you to the site. With one, it goes from you to the VPN server and then to the site. If the server is in your city, the detour is small. If it is on another continent, every request and every reply crosses an ocean it did not need to cross. This shows up as latency, the delay before anything starts, and it is the main thing you feel in calls and games.

2. The wrapper. Each packet you send is wrapped in another packet addressed to the VPN server. The wrapper takes space. WireGuard's author, Jason Donenfeld, broke the numbers down on the project's mailing list: an IP header of 20 bytes (IPv4) or 40 bytes (IPv6), 8 bytes of UDP header, 16 bytes of WireGuard header and a 16-byte authentication tag. That is 60 or 80 bytes out of a standard 1,500-byte packet, leaving 1,420 bytes for your data in the worst case. Call it four to five percent of each full-size packet spent on packaging. Older protocols spend more.

3. The math. Every packet is encrypted on the way out and decrypted on the way in. A recent phone or laptop does this faster than most home connections can deliver data, so you will not notice. A cheap router running the VPN for the whole house is a different story: its processor can become the limit long before your line does.

4. The server. A VPN server is shared. If many people are on the one you picked, or its own connection is busy, you get a smaller share.

Cause 1 affects delay. Causes 2, 3 and 4 affect throughput, the megabits per second on a speed test. They have different fixes, which is why one number from one speed test tells you so little.

How much does a VPN slow down your internet?

You will see figures like "10 to 20 percent" repeated across the web, almost never with a source. I am not going to add another one. Even the WireGuard project is blunt about its own published comparison, noting on its performance page: "These benchmarks are old, crusty, and not super well conducted."

The loss on your connection depends on things nobody else can measure for you:

• Your line speed. On a 50 Mbps line almost any modern device and server can keep up, so the loss is mostly the wrapper. On a multi-gigabit line the device or the server is more likely to be the limit, so the loss in percent can be larger.
• Distance to the server. Nearby costs milliseconds. Far away costs a lot of them.
• The protocol. Newer protocols such as WireGuard do less work per packet than older ones.
• Time of day. Evening load affects your provider and the VPN server alike.

So measure.

How to test how much your VPN slows you down

A single run with the VPN off and a single run with it on is not a measurement. Home connections vary from minute to minute by more than a VPN costs. This method controls for that.

Before you start

• Use one device. Plug it in with a cable if you can. If not, stay in one spot near the router.
• Close downloads, cloud backup, video and anything else using the connection. Ask the house for ten quiet minutes.
• In the VPN app, pick a server in or near your own city. You are measuring the VPN, not the distance to Tokyo.

Pick a test that lets you fix the far end

Speed test sites choose a test server near where you appear to be. Turn on a VPN and they quietly choose a different one, which ruins the comparison. Use one of these and keep the far end the same for every run:

• Cloudflare's speed test, which reports download, upload, latency and jitter on one page.
• The Ookla Speedtest command-line tool, where speedtest -L lists nearby servers and speedtest -s <id> runs against the one you choose.
• M-Lab's speed test, an open measurement project, as a second opinion.

Run it

1. VPN off. Run the test. Write down four numbers: download, upload, latency (ping) and jitter.
2. VPN on. Wait ten seconds. Run the same test. Write down the same four.
3. Repeat until you have five runs each way, alternating off and on. Alternating matters: if your provider slows down halfway through, both columns get hit equally.
4. For each number, take the median of the five, the middle value when sorted. Ignore the best and the worst.

Work out the loss

download loss % = (median off - median on) / median off x 100
added latency   = median on - median off, in milliseconds

Do the same for upload. You now have the real cost on your line, with your device, on that server.

Then test what you actually do

A speed test fills the pipe. Your evening does not. With the VPN on, start a video at the highest quality, join a call, load the game you play. If those work, the percentage is a curiosity.

What counts as too slow?

Compare what you have left with what the activity needs, not with the VPN-off number. The US Federal Communications Commission publishes a broadband speed guide with minimum download speeds per activity:

A 300 Mbps line that drops to 240 with a VPN on has lost a fifth of its speed and none of its usefulness. A 20 Mbps line that drops to 12 has a problem in a house with two people streaming.

For calls and games, look at latency and jitter instead. A small increase to a nearby server is hard to notice. A jump of a hundred milliseconds or more to a distant one is not.

Why is my VPN so slow, and how do I fix it?

Work down this list. Re-run the measurement after each change so you know what helped.

Two settings help when only part of your traffic needs protecting. Split tunneling lets you leave a game or a large download on your normal connection while everything else stays in the tunnel. And if your VPN offers multi-hop routing, keep it for the moments you want it: every extra server is another detour.

Can a VPN make your internet faster?

Occasionally, in one specific case. Some providers slow particular kinds of traffic, such as video from a certain service, while leaving the rest alone. A VPN hides what the traffic is, so that kind of selective slowing cannot target it.

You can test for this instead of guessing. Wehe, a research project from Northeastern University, replays an app's traffic pattern twice, once as itself and once with the content replaced by random bytes, and tells you whether your provider treated the two differently. If it did, a VPN may speed that app up. If your provider treats everything the same, a VPN can only add overhead.

Does post-quantum encryption make a VPN slower?

The extra work happens once, when the connection is set up, not on every packet. A post-quantum VPN uses a quantum-resistant algorithm to agree on a key, then encrypts your data with the same fast ciphers as before.

In Secria VPN, the post-quantum key is added at the handshake, not in the data path. Each connection gets a fresh ML-KEM-1024 key mixed into WireGuard, on by default. You can run the measurement above on it like on any other VPN.

FAQ

Does a VPN slow down Wi-Fi? It slows the connection running over the Wi-Fi by the same amount it would slow a wired one. It does not affect the Wi-Fi signal itself. If the slowdown is much worse on Wi-Fi than on a cable, the Wi-Fi is the weak link.

Does a VPN slow down your phone? It adds a little processor work and a little battery use. On a current phone the connection, not the processor, is the limit, so apps should feel the same.

Does a VPN use more data? Slightly. The wrapper on each packet counts toward your data allowance. With WireGuard that is roughly four to five percent on full-size packets, based on the header sizes published by its author.

Does a VPN increase ping? Yes, by the time it takes to reach the VPN server and come back. A server close to you, or close to the game server, keeps the increase small.

Is a free VPN slower than a paid one? Often, because free services limit speed or crowd many users onto few servers. Speed is the smaller concern with free VPNs. The bigger one is how they pay their bills, which we cover in are free VPNs safe.

Ten minutes with a notepad will tell you more about your VPN's speed than any review. If you want to run it on ours, Secria VPN comes with a 30-day money-back guarantee.