TL;DR: Every VPN slows your internet a little — that's the price of encryption and the extra hop your traffic takes through the VPN server. But the gap between "unnoticeable" and "unusable" comes down to your settings. With the modern WireGuard protocol and a nearby server, the loss is usually small and you won't feel it in everyday use. If your internet got noticeably slower with a VPN, the culprit is almost always server choice, the protocol, or a crowded free VPN — not the technology itself. Below: why speed drops, how to get it back in a few simple steps, and when a VPN actually speeds the internet up.
Why a VPN affects speed
A speed drop with a VPN on isn't a bug or, by itself, a sign of a bad service: it's a consequence of how the technology works. A VPN wraps your traffic in an encrypted tunnel and routes it through an intermediate server. Each of those steps costs a bit of time and computation, and together they add up to the difference you see on a speed test.
Your final VPN speed is set by four independent factors: your device's power (how fast it encrypts), the chosen protocol (how much computation and overhead per packet), distance to the server (physical latency), and that server's current load (how many others share the channel). A weak link in any of the four drags the whole chain down. So when things are slow, the right strategy is to walk through those four points one by one, not switch providers at random. The good news: almost all of these causes are manageable — you can't undo physics, but you can pick a nearby server, a fast protocol, and an uncrowded endpoint.
Speed and latency are different things
Before fixing a "slow VPN," understand exactly what's bothering you, because "slow" hides two different phenomena. Speed (bandwidth, megabits per second) determines how fast a large file, high-quality video, or a heavy page loads. Latency (ping, milliseconds) determines how quickly the server responds to your action — critical for online games, video calls, and interactive services.
A VPN affects both, but through different mechanisms. Bandwidth is mainly reduced by encryption and an overloaded server. Latency is increased primarily by distance to the server — the extra hop physically lengthens the signal's path. So if your games and calls are "laggy" while downloads are fine, fix latency — pick the nearest server. And if everything downloads slowly, dig into the protocol and server load. Understanding this difference saves time: you stop fiddling with random settings and hit the cause.
Encryption and CPU load
A VPN's main job is to encrypt your traffic. Every packet is encrypted on your device and decrypted on the VPN server, and vice versa. That takes CPU cycles. On most modern smartphones and laptops, chips hardware-accelerate common encryption algorithms, so the added latency is minimal — measured in fractions of a millisecond, with crypto running almost "for free." The real bottleneck is usually not encryption but distance to the server and its load. Encryption becomes a noticeable drag only in two cases: on genuinely old hardware, where the bare CPU does the crypto, and with a heavy protocol that maps poorly onto hardware acceleration.
The practical takeaway: if your device is weak, protocol choice affects speed even more than server choice. WireGuard's lightweight modern cryptography wins clearly here because it does less computation per packet. We covered how this protocol is built in our guide to the WireGuard protocol and its advantages. There's a flip side: a CPU busy with encryption uses more power, and on a smartphone that shows up as faster battery drain — so a light protocol improves both speed and battery life at once.
The extra route and distance to the server
Without a VPN, your traffic takes a short path: device → ISP → destination server. With a VPN, an intermediate link appears: device → ISP → VPN server → destination server. That detour lengthens the data path, and the key factor here is geography: the farther the VPN server physically sits, the longer the signal takes round-trip, and the higher your ping. The numbers are dictated by physics: light through fiber can't move faster than the speed of light, so connecting to a server on another continent inevitably adds noticeable latency.
Here lies a common beginner mistake: people habitually pick a server in a popular distant country (say, the US) even when they just need channel protection, not a specific American IP. As a result, traffic is routed across half the planet for nothing, and speed drops needlessly. Remember the rule: change the server's country only when you genuinely need an IP from that country (for example, to reach a regional service). In all other cases — security on public Wi-Fi, protection from ISP tracking, encrypting banking — pick the nearest server, and speed will barely suffer.
A few nuances. Sometimes a server labeled "Germany" physically sits in a neighboring country while the IP merely imitates a German geolocation — the real distance may be larger than the name implies. Also, the path to a server doesn't always match the straight line on a map: traffic follows a route built between ISPs and backbone carriers, and occasionally a server that seems farther actually gives lower latency because the route to it is better laid out. That's a reason not to guess but to test: run a measurement on the two or three nearest servers and keep the one that's actually faster.
VPN server load
Each VPN server handles many users at once, and its bandwidth is shared among everyone connected. If a server is overloaded, every client's speed drops — even with a nearby endpoint and a fast protocol. The problem is sharpest with free VPNs: they have few servers and thousands of users, so every endpoint is chronically overloaded, and the promised "unlimited speed" turns into a crawl. This is one of the hidden reasons a free VPN often costs more than a paid one. For more on what to look for when choosing, see our guide on how to choose a VPN in 2026.
A good paid service fights overload two ways: it keeps enough spare server capacity, and it shows each endpoint's current load or ping in the app so you can pick a free one. Some VPNs, including LiMP, connect you to the least-loaded server automatically. Load also has a daily rhythm: evening hours, when everyone watches video and browses social media, are peak load for both regular internet and VPNs. If your speed drops noticeably in the evening but is fine during the day, the cause is almost certainly hourly overload of popular servers — switch to a less obvious location in the same region. And remember: the app's load indicator is a useful guide, but the final word belongs to a real speed test here and now.
VPN protocol: why WireGuard is faster
A protocol is the set of rules used to build the encrypted tunnel, and it affects speed as much as anything else. Different protocols process data differently and load the CPU differently.
- OpenVPN — a reliable, time-tested protocol, but it runs in user space and processes each packet through the application. That creates extra overhead, and on fast connections it loses noticeable speed.
- IKEv2/IPSec — a middle ground: faster than OpenVPN, especially good on mobile when switching between Wi-Fi and cellular, but harder to configure.
- WireGuard — a modern protocol with compact, efficient cryptography. On desktop systems it runs at the OS kernel level; on mobile (for example, on iOS) it runs as a userspace network extension. Either way, in practice it's usually noticeably faster than OpenVPN and uses fewer resources.
If your provider offers a protocol choice, for speed you should almost always pick WireGuard. The exception is rare cases where a network blocks non-standard protocols and you have to "disguise" as regular HTTPS traffic via OpenVPN on port 443.
Why does WireGuard win? First, it has a very compact codebase and a fixed set of modern cryptographic algorithms: it doesn't need to negotiate per connection which cipher to use, as the flexible but bulky OpenVPN does. Second, where WireGuard runs in the kernel, packets don't have to "switch" between the app and the system — and those switches are what eat time in older protocols. Less code and fewer "negotiations" mean less latency. That said, a newer protocol isn't always automatically the best for every task: IKEv2, for example, has historically been very good on mobile with frequent network changes. But for speed in the general case, WireGuard is today's optimal balance of performance, efficiency, and encryption strength.
What speed loss counts as normal
To know whether you have a problem, you need a baseline. First, measure your speed without the VPN on any speed test, then with the VPN on the nearest server, and compare. On WireGuard and a nearby server the drop is usually small and unnoticeable in everyday use: video, messengers, and browsing work as if the VPN weren't there. A noticeable loss on demanding tasks is a reason to try another server or switch to WireGuard. And if the internet is several times slower with the VPN — that's a problem: the cause is almost always a distant or overloaded server, a slow protocol, or your ISP itself.
Two important caveats. A speed test with a VPN always measures speed to a specific test server, and the result depends on where it sits relative to your VPN endpoint — run several tests on different servers so the picture is honest. And judge the loss not in absolute megabits but relative to your tasks: on a fast plan, even a noticeable "loss" leaves you plenty for 4K video and games. The only right question isn't "how many megabits did I lose" but "is the remaining speed enough for what I'm doing."
What matters for streaming, gaming, and calls
Different tasks place different demands on the connection, and the "sufficient" speed for them varies a lot.
- Video streaming. Stable bandwidth matters, not record bandwidth. Full HD needs a few megabits, 4K needs a few dozen. The key is stability without stutters, which a nearby, uncrowded server provides.
- Online games. Latency (ping) is critical, not bandwidth — little data is transferred, but it must arrive instantly. The nearest server with a short route matters most to gamers; a distant one adds tens of milliseconds, felt as "lag."
- Video calls. They need moderate bandwidth, low latency, and stability without packet loss. WireGuard with a nearby server handles this excellently; problems start on distant or overloaded endpoints.
- Downloads and cloud. Only bandwidth matters, latency is almost irrelevant. This is the ideal candidate to route around the VPN via split tunneling, if the files aren't sensitive.
Bottom line: before getting upset about "speed loss," ask whether it hampers your actual task. Often the remaining speed is more than enough, and the feeling of "slowness" comes not from bandwidth but from latency — fixed by switching to the nearest server.
How to get your speed back: a step-by-step checklist
Work through the list top to bottom — the problem is usually solved in the first two or three points.
- Pick the nearest server. If you don't need an IP in a specific country, connect as close to yourself as possible. This gives the biggest speed gain and the largest ping reduction.
- Switch to WireGuard. If the app lets you choose a protocol, set WireGuard. On weak devices and fast connections the difference is especially noticeable.
- Try a different server in the same region. Even neighboring servers vary in load — check the load or ping indicator and pick a free one.
- Set up split tunneling. Route only what needs protecting (browser, banking, messengers) through the VPN, and send heavy, non-sensitive traffic — updates, cloud sync — directly.
- Measure your baseline speed without the VPN. Sometimes the connection itself is slow, not the VPN. If speed is low without the VPN, the problem is your ISP and changing tunnel settings is pointless.
- Update the VPN app. An outdated version can run slower than a new one for no visible reason.
- Reboot your device and router. Stuck network processes and an overflowing router cache are often the real source of the slowdown.
- Drop the free VPN. If you've hit a speed ceiling on a free service, no settings will help — there simply aren't enough servers for everyone.
Work through them in order and check the result after each step — that way you'll pinpoint the weak link instead of changing everything blindly. In most cases the first two points are enough: the nearest server plus WireGuard bring speed back to comfortable.
When a VPN actually speeds the internet up
It sounds paradoxical, but sometimes a VPN can increase your real speed. This happens when your ISP artificially throttles certain traffic types — video streaming, torrents, or specific services. The ISP identifies the traffic type using deep packet inspection and slows down exactly that. Reasons vary: saving on backbone capacity at peak hours, commercial arrangements, or an attempt to relieve the network at the expense of the heaviest services. For you the result is the same: you pay for a plan but don't get the promised speed on specific services.
A VPN encrypts all traffic as one stream, and the ISP can no longer see what's inside the tunnel: to it, it's just a uniform encrypted flow. Since it can't identify the traffic type, it can't selectively slow it down — video that used to "buffer" plays smoothly again. How do you tell throttling from ordinary congestion? The signs are distinctive: a strictly defined traffic type suffers (for example, only video or one streaming service), the effect is stable day after day, while the channel's speed on neutral speed tests is normal. If you turn on the VPN and the problem disappears specifically on that service, it was almost certainly throttling. If the VPN changes nothing or makes it worse, the issue isn't your ISP but congestion or distance to the server.
Speed and security: a false trade-off
A common beginner fear: "if I want speed, I'll have to sacrifice protection, and vice versa." With modern protocols this is a false dilemma. WireGuard delivers both strong encryption and high speed at once — you don't have to choose. The myth lingers from the early days of VPNs, when encryption genuinely was a heavy load on weak processors; hardware has moved on since, and the "cost" of encryption has shrunk to fractions of a millisecond. Chasing "weaker encryption for megabits" is pointless: the speed difference is negligible, while the privacy risk is real.
What's more, trying to "save" on protection for the sake of speed often backfires. For example, disabling the emergency disconnect (kill switch) won't add speed, but if the tunnel drops, your traffic will leak onto the open network without you noticing. It's smarter to keep protection on and configure split tunneling instead: heavy, non-sensitive traffic goes directly and fast, while important traffic goes through the secure tunnel. For more on the safety net against drops, see our guide on what a VPN kill switch is, and on protection in untrusted networks, our guide to public Wi-Fi security.
What does NOT affect VPN speed, though many think it does
- "The pricier the plan, the higher the speed." Price alone doesn't buy speed. What matters is infrastructure (the number and quality of servers) and the protocol, not the price tag.
- "Encryption key length directly cuts speed." On modern devices the speed difference between encryption levels is negligible — hardware acceleration absorbs it.
- "A VPN is always slower, and nothing can be done." A half-truth: slower, yes, but by how much depends entirely on your settings.
- "Restarting the app once an hour speeds up the VPN." No. If the connection is stable, restarts do nothing; a restart only helps with a genuinely stuck connection.
The takeaway is simple: instead of ritual actions, hit the real four factors — device, protocol, distance, and server load. Everything else is noise.
Conclusion
A VPN inevitably slows the internet a little — that's physics and cryptography, not a flaw of a specific service. But the right settings keep the loss minimal: pick the nearest server, use WireGuard, stay off crowded free endpoints, and enable split tunneling when needed. A "slow VPN" almost always means a "misconfigured VPN" or a "bad service," not a fundamental problem with the technology. Walk through the four manageable factors — device, protocol, distance, and server load — find your optimal combination, and you'll forget about "slowdowns" for a long time.
If you're looking for a VPN that doesn't turn fast internet slow, LiMP runs on WireGuard by default, keeps spare server capacity, and connects you to the nearest free endpoint automatically. It's iOS and Android, 100 ₽/month, and no logs of your activity. Terms and sign-up are on the pricing page.
Frequently asked questions
Why does my VPN speed drop only in the evenings?
Most often it's hourly overload of popular servers: in the evening everyone watches video and browses social media at once, and the channel is shared among more people. During the day the same servers are freer. The fix is to switch in the evening to a less obvious location in the same region or to a server with a lower load indicator.
How do I measure exactly how much a VPN slows my internet?
Run a speed test without the VPN, then with the VPN on the nearest server, and compare as a percentage of your baseline. Run several tests, since the result also depends on the test server. Judge not the absolute megabits but whether the remaining speed is enough for your tasks.
Should I switch internet providers if everything is slow with a VPN?
First check your baseline speed without the VPN: if everything is fast without it, the problem is in the VPN settings (server, protocol, overload), not the provider, and switching is pointless. If speed is low even without the VPN, then the issue really is your provider's link.
Does a VPN affect Wi-Fi speed more than a wired connection?
The VPN itself works the same in both cases, but Wi-Fi adds its own loss and latency on top of the VPN, especially far from the router or with interference. On a computer, where possible, a wired connection removes that variable and helps tell whether the VPN or your home network is at fault.
Does a VPN slow speed on a phone more than on a computer?
Not necessarily. Modern smartphones hardware-accelerate encryption, so the difference is small. A noticeable slowdown on a phone is more often due to a weak network signal, switching between Wi-Fi and cellular, or an old device model, rather than the mobile platform itself.
Can I use a VPN all the time without losing speed?
Yes. With the nearest server and WireGuard, speed loss is practically unnoticeable in everyday use, so a VPN can stay on all the time. If something is slow during constant use, the cause is a specific setting (a distant server or a crowded endpoint), not the fact that the VPN has been on for a long time.