A VPS in Amsterdam is the default choice for a site whose visitors are in the Netherlands, Belgium or the UK. On 13 November 2026 we measured round trips from our Amsterdam node of 1.3ms to a KPN network address, 6 to 8ms to the two largest Belgian ISPs and under 10ms to London. The US East Coast came in at 85 to 90ms.
The rest of this guide shows how we measured those numbers and what they mean for page load. It also covers how Amsterdam compares with Frankfurt and London, what Dutch data residency does and does not settle, and what to check on any Dutch VPS offer before you pay for it.
Why so much Dutch hosting ends up in Amsterdam
The Netherlands is a small country with a lot of internet. Most of it meets in and around Amsterdam, at AMS-IX and the smaller exchanges and carrier hotels nearby. The large Dutch ISPs hand traffic over there, and so do many of the European and transatlantic networks your visitors might be on.
The practical effect is that a server in Amsterdam is close to everyone in the country, not only to people in the capital. A visitor in Groningen or Maastricht reaches it through their ISP’s own backbone and a short hop at the exchange. There is rarely a reason to look for a VPS in Rotterdam or Eindhoven unless you need a second site for redundancy.
Our Amsterdam VPS runs on Leaseweb’s Netherlands network (AS60781), which is connected at AMS-IX. Every figure below was taken from that node.
What we measured, and how
We sent 30 ICMP echo requests to each destination from the Amsterdam node, 0.2 seconds apart, and recorded the average round trip. Every destination in the table answered all 30, with no loss. For Dutch and Belgian visitors we pinged DNS resolvers that sit inside each ISP’s network, checked against the RIPE database, because a resolver is about as close to that ISP’s customers as a public address gets. The other destinations are our own nodes.
| Destination | What we pinged | Round trip, 13 Nov 2026 |
|---|---|---|
| Netherlands | KPN resolver, 195.121.1.34 | 1.3 ms |
| Belgium | Proximus resolver, 195.238.2.21 | 6.3 ms |
| Belgium | Telenet resolver, 195.130.131.4 | 8.0 ms |
| United Kingdom | Our London node | 7.6 ms (IPv6), 9.6 ms (IPv4) |
| Germany | Our Nuremberg node | 12.5 ms |
| US East Coast | Our Washington DC node | 85 to 90 ms |
| US West Coast | Our San Francisco node | 110.9 ms |
| Singapore | Our Singapore node | 199 to 212 ms |
A few notes on reading this table, because round-trip figures are easy to over-read.
Ziggo, the other big Dutch consumer ISP, is missing. Its resolvers did not answer pings or connections from our node, so we could not measure it the same way. We expect it to land close to KPN, since both hand traffic over in Amsterdam, but we have not measured it, so it is not in the table.
London answered two milliseconds faster over IPv6 than over IPv4. The two protocols probably took slightly different routes. Washington DC moved between 85 and 90ms from one run to the next. That kind of drift is normal and is why we give a range.
Singapore is the outlier. In October the same route measured 157ms. On 13 November it sat between 199 and 212ms over several runs, so the path between Europe and Asia had changed. If a large share of your visitors are in Asia, Amsterdam is the wrong place for your server whichever number you take.
These are round trips on networks we run or can see. A visitor’s own latency adds their ISP’s access network, which on fibre is often a millisecond or two and on mobile can be much more.
What those milliseconds mean for a page load
Latency matters because a browser needs several round trips before it receives anything. On a first visit over HTTPS with TLS 1.3, the TCP handshake takes one round trip, the TLS handshake another, and the request for the page a third. Three round trips before the first byte is a fair rule of thumb.
At Dutch distances that is about 4ms. For a visitor in London it is about 25 to 30ms, and for one in Washington DC about 260ms. Repeat visits and HTTP/2 connection reuse cut the count, and a CDN takes images and scripts off the long path. The HTML from your application still comes from the server, so this part of the delay stays.
Server location is only part of time to first byte, and often the smaller part. A slow database query or an uncached WordPress page can add 500ms at the origin, which dwarfs anything in the table above. We go through the other parts in what TTFB is and why your hosting provider controls it.
Amsterdam, Frankfurt or London?
These three are the usual shortlist for a European VPS, and each has a clear case.
Amsterdam suits sites whose visitors are in the Netherlands, Belgium and the UK, or spread across Western Europe with some in North America. Washington DC is 85 to 90ms away, close enough for one server to serve both.
Frankfurt is the better choice when most visitors are in Germany, Austria or Switzerland. DE-CIX in Frankfurt is where much of Central European traffic is exchanged. We sell VPS in Frankfurt at the same price as Amsterdam. We have no test node there yet, so we cannot give you our own Frankfurt figures. Our Nuremberg node, which is further east, is 12.5ms from Amsterdam.
London makes sense when your visitors are almost all in the UK, or when a contract or regulator asks for data to stay in the UK itself. For speed alone, our measurements suggest a British audience would struggle to tell London from Amsterdam.
If your traffic is genuinely spread across Europe, do not agonise over it. From Amsterdam, London is under 10ms away and Nuremberg 12.5ms, so the gap between any two of these cities is a matter of milliseconds, and any one of them is a sound choice.
Data residency: what a Dutch VPS settles, and what it does not
Hosting in the Netherlands puts the server and the data on its disk in an EU member state. The Dutch Data Protection Authority, the Autoriteit Persoonsgegevens, supervises how it is processed. For many businesses that answers the first question a customer or auditor asks: where is the data?
It does not finish the job. Under GDPR Article 28 you need a written Data Processing Agreement with any provider that processes personal data for you. Ours is published at /dpa, and you should expect the same from any host you consider. Residency also covers only the server. If your backups go to a storage service in another country, or your contact form sends mail through a provider elsewhere, that data travels with it. Map those flows before you tell anyone everything stays in the Netherlands.
This is general information, not legal advice. If you work under sector rules such as health or finance, ask your compliance lead which locations and certifications they need, then choose the location.
What to check before you buy a VPS in the Netherlands
Prices for Dutch VPS plans vary a lot, and the headline number hides most of the differences. These are the points that separate one offer from another.
- Virtualisation type. KVM gives you your own kernel and reserved RAM. Container-based plans share the host kernel, which limits what you can run.
- Storage. NVMe is now normal. A plan that does not say which kind of disk it uses usually means SATA SSD or slower.
- Traffic and port speed together. A 10 Gbps port with a small monthly allowance and a 1 Gbps port with 30 TB are very different plans. Check what happens when you pass the allowance: throttling, a bill, or suspension.
- The network, by name. Ask which network and AS number the server sits on, then test it. A provider that will not tell you is asking you to buy blind.
- IPv6. A routed /64 should be included at no extra cost. Some budget providers still leave it out or charge for it, so ask before you order.
- DDoS protection. Find out whether it is included, how much it absorbs and what happens to your server during an attack. Some providers simply null-route the address.
- Operating systems. Check that the distribution you use is offered in a supported release. Not every Dutch VPS offers Windows; ours does not.
- Billing currency and VAT. Euro billing avoids conversion fees for Dutch and Belgian businesses. A VAT-registered business customer should be able to supply its VAT number.
Our Amsterdam VPS plans
Every plan on our VPS line can be deployed in Amsterdam with no location surcharge. All of them run on AMD EPYC processors with NVMe storage, include 30 TB of traffic on a 10 Gbps port, a free IPv4 address and a /64 of IPv6, and come with DDoS protection.
| Plan | vCPU | RAM | NVMe storage |
|---|---|---|---|
| VPS 400 | 4 | 6 GB | 100 GB |
| VPS 600 | 6 | 16 GB | 200 GB |
| VPS 800 | 8 | 24 GB | 300 GB |
| VPS 1200 | 12 | 48 GB | 400 GB |
| VPS 1600 | 16 | 64 GB | 500 GB |
| VPS 2400 | 24 | 120 GB | 600 GB |
You choose Ubuntu, Debian, AlmaLinux, Rocky Linux, Fedora or FreeBSD at checkout, with full root access. cPanel or Plesk can be added if you want a control panel. Current prices for each plan are on the Amsterdam VPS page, where the order button opens the cart with Amsterdam already selected. To compare every location side by side, see the full VPS page.
If you are still deciding whether you need a VPS at all, shared hosting vs VPS: when to upgrade walks through the signs.
Test it yourself
Our table tells you about our network. Yours will tell you about your visitors. Before you order, run this from your office, a customer’s connection or a server you already rent:
ping 95.211.93.118
mtr 95.211.93.118
The browser speed test at st-ams.gozen.host measures the same route from inside a web page, and the network page has test addresses for every other location. If Amsterdam comes out well for the people who matter to you, that is a better reason to choose it than any table we could publish.
Frequently asked questions
Is Amsterdam a good location for a VPS?
For visitors in the Netherlands, Belgium and the UK, yes. Measured from our Amsterdam node on 13 November 2026, a KPN network address answered in 1.3ms, Belgian ISPs Proximus and Telenet in 6 to 8ms, and London in under 10ms. The US East Coast is 85 to 90ms away, which is workable for a site that serves both sides of the Atlantic from one server.
Should I choose Amsterdam or Frankfurt for a European VPS?
Amsterdam is usually the shorter path to the Netherlands, Belgium, the UK and the US East Coast. Frankfurt sits closer to Germany, Austria and Switzerland. If most of your visitors are German-speaking, choose Frankfurt. Otherwise Amsterdam is the safer default. On our VPS line both cost the same.
Does a VPS in the Netherlands keep my data in the EU?
The server and the data on its disk stay in the Amsterdam facility, under EU law and the supervision of the Dutch Data Protection Authority (Autoriteit Persoonsgegevens). That covers where the data is stored. Under GDPR you also need a Data Processing Agreement with your host, and you remain responsible for any service that copies the data elsewhere, such as an external backup or email provider.
How fast is an Amsterdam VPS for visitors in the UK?
Our London node answered in 7.6ms over IPv6 and 9.6ms over IPv4 from Amsterdam. A first visit needs about three round trips before the first byte arrives, so the network adds roughly 25 to 30ms. For most UK sites that is not noticeable next to the time the application itself takes.
Can I test the latency to your Amsterdam server before buying?
Yes. Run ping or mtr against 95.211.93.118 from the network your visitors use, or open the browser speed test at st-ams.gozen.host. Your own numbers matter more than ours, because they include your ISP and your last mile.
Test the route before you order
Ping 95.211.93.118 from your office or from a server you already run. If the numbers suit your visitors, the Amsterdam page lists every plan with the location already selected in the cart.
Andreas writes about hosting infrastructure, performance, and migrations at GOZEN HOST LLC, a Top 25 WordPress Hosting Provider for 2026 (HostAdvice).