Why the nearest VPN server is not always the quickest
The closest VPN server on a map can be slower than one further away, because internet routes, peering and server load matter more than kilometres.
The nearest VPN server on a map is often the quickest, but not reliably, and the gap can be large. Internet traffic does not follow geography; it follows the agreements between networks, so a server 300 km away on a direct route can answer sooner than one 50 km away reached through a distant exchange. On top of that, a nearby server that is busy will feel slower than a distant one with capacity to spare. The only trustworthy way to know is to measure the round trip from your own phone.
Distance sets a floor, not the answer
Distance does matter. A signal in fibre covers roughly 100 km per millisecond, so each 100 km between you and the server adds around 1 ms in each direction. That is a floor: no route can beat it. But it is only a floor. Real paths add routing on top, and the routing is where the surprises live. The latency versus bandwidth explainer covers how the two combine.
Routes do not follow the map
Your carrier hands traffic to other networks at exchange points, and it chooses those hand-offs by cost and agreement, not by distance. Two common effects follow.
Traffic between neighbours often goes via a hub. A packet from one city to a nearby city in a different country may travel to a large exchange in a third country and back, because both carriers peer there and nowhere closer. The map says 200 km; the route says 1,500 km.
Some carriers route almost everything through one place. A mobile carrier that connects to the wider internet mainly through one exchange will show low latency to anything near that exchange and higher latency to everything else, regardless of where the phone is. From that phone, a server beside the exchange is effectively the nearest one.
Neither of these is visible in the app's location list. Both are visible in a measured round trip.
The server itself has its own route out
Once your traffic reaches the VPN server, it still has to get to the site you are using, and that leg depends on the server's hosting network. A server in a well-connected data centre with direct links to the large content networks will serve pages quickly; a server in a poorly connected one will not, no matter how close it is to you. This is also why a VPN occasionally seems quicker than no VPN: the server's route to a site is better than your carrier's.
Load can dwarf distance
A VPN server shares one uplink and one processor among everyone connected to it. When it is near capacity, each session's share of bandwidth falls and the round trip becomes jittery, because packets queue. A busy server 50 km away can behave worse than a quiet server 800 km away for anything heavier than messaging. What server load is goes through how it shows up and why an app should rank by it.
What the measurements look like in practice
The table shows the kind of pattern you see when the map and the network disagree. The numbers are illustrative of the shape, not measurements of any particular city.
| Server | Map distance | Round trip you might measure | Why |
|---|---|---|---|
| Just across a border | Short | Higher than expected | Carriers peer at a distant hub |
| A large hub city further away | Medium | Lower than expected | Direct route through the hub |
| Very close, but busy | Shortest | Low but jittery, slow to move data | Queued packets, shared uplink |
| Far away, quiet | Long | High, but steady and roomy | Distance floor, no queueing |
For browsing and messaging the second row usually wins. For a large download the fourth row can beat the third. For a call or a game, the lowest steady round trip wins, which is often the second row again.
How to choose
Ignore the map and look at two figures: the round trip measured from your phone, and how much headroom the server has. Choose the lowest steady round trip for real-time use, and accept a slightly higher round trip for more headroom when you are moving a lot of data. If the app only shows a bar icon or a country flag, it is asking you to guess, and how a good VPN app picks a server explains what it should be doing instead.
Measure again when things change. A route that was good at home can be poor on a different carrier, and a city that was quiet in the morning can be full by evening. The figure only means something if it was taken from your phone, on the network you are on, a moment ago.
Culvert VPN measures a live round trip from your phone to each city, shows real headroom beside it, and when you tap a country connects you to the city with the lowest measured round trip and most headroom rather than the closest one on the map; it is free on Google Play.
Questions people also ask
How do I find the quickest server without trying them all?
Use an app that measures the round trip from your phone to each city before you connect and shows the result as a number. That measurement already includes the routing, so you do not have to reason about it.
Does the quickest city change over the day?
Yes, and sometimes by a lot. Load rises and falls, and carriers change routes. A city that measured 40 ms in the morning can measure 90 ms in the evening, which is why the figure should be live rather than stored.
Should I pick a server in my own country?
Only if it measures well. A server in a neighbouring country on a direct route with plenty of headroom is often a better choice than one at home behind a congested exchange.
All 12 guides in Speed, battery and reliability · Every guide