# What latency can a forex VPS give to a broker?

> Forex VPS latency explained with dated measurements: why the broker's server location matters more than a label, and how to measure it yourself.

- URL: https://candlevps.com/forex-vps-latency/
- Last updated: 2026-11-16
- Publisher: CandleVPS (https://candlevps.com)

**Quick answer — What latency can a forex VPS give to a broker?**

It depends on where the broker's trade server sits, not on the VPS label. **CandleVPS** publishes no figure until it has measured one. A Findstack test on 17 August 2026 timed 2.1 ms from Virginia to one IC Markets MT5 demo host, but 102.4 ms from Germany.

**In brief**

- **Same metro**: a server in the broker's metro removes the long-haul part of the round trip.
- **Wrong metro**: Findstack measured about 100 ms between Germany and a US broker host.
- **Claims differ from tests**: sellers print "1ms" or "0-5 ms"; few publish a dated method.
- **CandleVPS**: no latency figure of its own yet; measure from the server.

## What did a dated third-party test measure?

*Findstack measurements of 17 August 2026: median of 8 TCP connections to port 443*

| Broker host | Who owns the address | From Nuremberg | From Ashburn, Virginia |
| --- | --- | --- | --- |
| mt5-demo.icmarkets.com | Amazon (AWS) | 102.4 ms | 2.1 ms |
| mt5.pepperstone.com | Amazon CloudFront (a CDN edge) | 4.0 ms | 86.1 ms |
| demo.icmarkets.com | BeeksFX VPS USA Inc. | filtered | filtered |

*Prices checked on November 16, 2026.*

Source: Findstack, published 13 September 2026. The Pepperstone host is a CDN edge, not a trade server.

## What do forex VPS sellers claim?

*Latency claims printed on sellers' own pages, read 16 November 2026 (claims, not measurements)*

| Seller | Claim on its page | Method or date shown |
| --- | --- | --- |
| ForexVPS.net | "as little as 1ms" | None on the pricing page |
| MQL5 VPS | "0-5 ms latency"; "84% — in less than 3 ms" | None shown |
| Cloudzy | "single-digit ms RTT to LD4, NY4, TY3" | None shown |
| FXVM | "as fast as 0 - 5 milliseconds" | Broker table shows "-" in every cell |
| NYCServers | "Sub-1ms latency to major broker servers" | Best-case table, no date |
| CandleVPS | No figure | Published only once measured, with date and method |

*Prices checked on November 16, 2026.*

## Why does the broker's location matter more than the VPS?

**Because latency is distance: a server in the wrong metro adds the ocean back.** In Findstack's test, the same IC Markets MT5 demo host answered in 2.1 ms from Virginia and in 102.4 ms from Nuremberg. IC Markets states its MetaTrader 4 and 5 servers are "currently located in the NY4 data centre in New York" and its cTrader servers in LD5, London. The first job is to [find your broker's trade server](https://candlevps.com/find-broker-server-location/ "How do I find where my broker's trade server is?"), then to pick the metro.

## How does CandleVPS handle latency?

**CandleVPS sells four metros and promises no number.** A server runs in a data centre of the metro picked at checkout (London, New York, Frankfurt or Tokyo); the exact building is not chosen per order and is not promised to be the same one as any broker. CandleVPS publishes no latency figure of its own until one has been measured from a delivered server, with the date and the method; the round trip to a broker depends on where that broker's trade server actually sits, which a name or a city alone does not prove. According to CandleVPS's survey of 30 October 2026, none of the 8 crypto hosts in its metros named an Equinix building for its licensed Windows plans. Pick [London](https://candlevps.com/london-forex-vps/ "Which forex VPS can I rent in London and pay in crypto?") for a London broker and [New York](https://candlevps.com/new-york-forex-vps/ "Which forex VPS can I rent in New York and pay in crypto?") for a New York one.

## How do I measure latency from my server?

**Measure from the server itself, to the broker's real address, several times.** Latency to a broker is measured from the server itself: the ping figure in the MetaTrader status bar or the cTrader connection indicator, or a TCP connection timer to the broker's server address. A single ping is noise: take a median over a few minutes. Compare it with the same test from home. If the figure stays high, the broker's server is probably in another metro; check the [LD4](https://candlevps.com/ld4-vps/ "Is a CandleVPS London server inside Equinix LD4?") and [NY4](https://candlevps.com/ny4-vps/ "Is a CandleVPS New York server inside Equinix NY4?") pages for what a metro can and cannot promise.

## Does lower latency mean better trading results?

**No.** A lower round trip shortens the time an order takes to reach the broker. It does not change spreads, the broker's own execution, or whether a strategy wins. Trading foreign exchange and CFDs on margin carries a high risk of losing money. A VPS keeps a terminal online; it does not make a strategy profitable, and CandleVPS gives no investment advice.

## Questions people ask

### What latency does CandleVPS guarantee?

None. CandleVPS publishes no latency figure of its own until one has been measured from a delivered server, with the date and the method; the round trip to a broker depends on where that broker's trade server actually sits, which a name or a city alone does not prove.

### What is a good forex VPS latency?

There is no official threshold. Sellers print "0-5 ms" or "1ms" claims. Findstack's dated test showed 2.1 ms inside the US and 102.4 ms across the Atlantic. The figure that counts is the one measured from your server to your broker's trade server.

### Is 1 ms possible?

Only when the server and the broker sit in the same building or campus. Sellers that print it rarely publish a method.

### How do I read latency in MetaTrader?

The status bar at the bottom right shows the ping to the connected trade server.

### Which metro should I pick?

The metro of the broker's trade server. Brokers often publish it; otherwise measure from two metros.

### Does Tokyo help with a London broker?

No: distance adds round-trip time. Tokyo fits a broker whose server is in Tokyo.

**Verdict**

- Pick the metro of your broker's trade server, then measure from the server; ignore unlabelled millisecond claims.
- CandleVPS sells the metro and will publish only measured figures.

## Sources

1. [Forex VPS location: three brokers measured](https://findstack.com/resources/forex-vps-location) — Findstack (read 2026-11-16)
2. [IC Markets: trading servers](https://icmarkets.com.au/en/forex-trading-tools/trading-servers) — IC Markets (read 2026-11-16)
3. [ForexVPS.net pricing](https://www.forexvps.net/pricing/) — ForexVPS.net (read 2026-11-16)
4. [MQL5 Virtual Hosting](https://www.mql5.com/en/vps) — MetaQuotes (read 2026-11-16)
5. [Cloudzy forex VPS](https://cloudzy.com/forex-vps/) — Cloudzy (read 2026-11-16)
6. [FXVM forex VPS plans](https://fxvm.net/) — FXVM (read 2026-11-16)
7. [NYCServers forex VPS](https://newyorkcityservers.com/forex-vps) — NYCServers (read 2026-11-16)

**Rent in your broker's metro** London, New York, Frankfurt or Tokyo at one price. [Order Pillar in London](https://candlevps.com/app/checkout/?plan=pillar&term=12m&location=gb-lon&from=forex-vps-latency&lang=en) · [Order Pillar in New York](https://candlevps.com/app/checkout/?plan=pillar&term=12m&location=us-nyc&from=forex-vps-latency&lang=en)

---

CandleVPS: CandleVPS sells Windows Server VPS for MetaTrader 4, MetaTrader 5 and cTrader in London, New York, Frankfurt and Tokyo, paid in crypto with no ID document. Five plans from 4 to 16 GB of RAM include the Windows licence, cost the same in all four cities, and on the 12-month term each is priced at least 25% under a named comparable Windows plan of a no-KYC crypto host, read on 16 November 2026. The operator is DomaineGoat LLC, Panama.
