Free managed migration on every plan — we move your sites the same day. All systems operational +91 74091 34231

ReportMeasured 25 September 2026

Server latency from India, city by city.

We measured round trips from 28 home and office connections in 20 Indian towns and cities to servers in seven regions. A server in Mumbai answered in a median 35 ms. Singapore took 73 ms, Europe 152 ms and the United States 247 ms.

35 ms
Mumbai, median
73 ms
Singapore, median
152 ms
London, median
247 ms
Virginia (US East), median

01By region

Distance, in round trips.

  • For 28 of 28 connections the nearest region was in India, Mumbai or Hyderabad.
  • Singapore was 2.1 times the Mumbai figure, and added a median 40 ms to every round trip over the nearer Indian region.
  • Europe was 4.4 times and the United States 7.2 times as far as Mumbai.
Mumbai, India35 ms · middle half 24–45
Hyderabad, India37 ms · middle half 24–47
Singapore73 ms · middle half 54–82
Frankfurt, Germany155 ms · middle half 146–166
London, United Kingdom152 ms · middle half 142–162
Virginia, United States247 ms · middle half 234–273
California, United States255 ms · middle half 238–264

Median round trip across all 28 connections. "Middle half" is the range the central 50% of connections fell in.

02By city

City by city.

Median round trip in milliseconds. The shaded cell is each place's nearest region. Places are named as Globalping's geolocation reports them; most have one or two probes, so read single rows as examples rather than averages.

Scroll the table sideways to compare

From MumbaiHyderabadSingaporeFrankfurtLondonVirginiaCalifornia
Agra 29 38 80 160 153 275 261
Aizawl, Mizoram (3) 54 59 82 184 177 279 252
Barddhaman, West Bengal (3) 43 36 73 155 156 243 271
Bengaluru (2) 25 23 46 154 156 247 225
Bishnupur 30 46 80 163 155 273 258
Chennai 24 15 35 146 140 234 234
Delhi (2) 25 31 68 143 138 227 266
Edattala, Kerala 47 38 53 194 167 281 234
Guwahati 51 42 94 184 180 255 267
Hyderabad 20 18 48 137 135 215 229
Jaipur 34 54 84 159 150 274 249
Jammu 44 51 104 174 169 288 277
Kanayannur, Kerala 45 36 51 164 144 246 244
Kolkata 36 37 73 155 153 241 275
Mumbai (2) 5 19 59 135 128 230 248
New Delhi 17 24 45 149 147 249 225
Prayagraj, Uttar Pradesh (2) 35 46 84 155 152 233 263
Pune 4 16 61 136 130 245 239
Shillong, Meghalaya 50 51 94 184 175 260 265
Thane 9 24 70 135 137 212 255

03What it costs

Three round trips before the first byte.

A browser opening a new HTTPS connection over TCP and TLS 1.3 needs three round trips before the first byte of a page arrives: one to open the connection, one to agree the encryption and one for the request itself. HTTP/3 saves one; older TLS 1.2 needs one more. That is network time alone, before the server does any work.

Mumbai104 ms
Hyderabad110 ms
Singapore218 ms
Frankfurt464 ms
London457 ms
Virginia741 ms
California765 ms

A CDN such as Cloudflare hides the distance for files it has cached, like images, stylesheets and scripts, by answering from a city near the visitor. It cannot for anything built per visitor: uncached pages, logins, carts, checkouts, dashboards and API calls still travel to your server and back.

04Method

How it was measured.

We sell hosting in India, so we have an interest in the answer. That is why the test targets neutral cloud endpoints rather than our own servers, runs on a public network anyone can use, and publishes every probe's result. The per-city figures on our product pages come from our own monitoring of our servers, measured separately, so the two sets are not directly comparable.

Probes. Globalping, the free, open measurement network run by jsDelivr, has small probes on ordinary internet connections. We asked for every probe in India that Globalping classifies as a consumer or business ISP network rather than a data centre: 38 answered the call. We then left out 7 that sit on cloud or hosting networks despite that label (Oracle Cloud, HostDzire, Good Domain Registry, HostStack) and 3 that did not answer for every region, leaving 28 connections on 14 networks in 20 places.

Targets. One stable public endpoint per region, AWS's regional DynamoDB hosts, the same ones cloudping.info uses. They stand in for "a server in that city"; this is a test of distance, not of AWS or of any host.

  • Mumbai: dynamodb.ap-south-1.amazonaws.com
  • Hyderabad: dynamodb.ap-south-2.amazonaws.com
  • Singapore: dynamodb.ap-southeast-1.amazonaws.com
  • Frankfurt: dynamodb.eu-central-1.amazonaws.com
  • London: dynamodb.eu-west-2.amazonaws.com
  • Virginia: dynamodb.us-east-1.amazonaws.com
  • California: dynamodb.us-west-1.amazonaws.com

Measurement. Each probe opened 16 TCP connections to port 443 of each endpoint and timed each handshake, one network round trip. A probe's figure is the median of its 16; a region's figure is the median across probes. Every region was measured from the same probes, on 25 September 2026 between 06:03–06:04 IST.

Limits. This is one snapshot, taken early in the Indian morning, when networks are quiet; evening peaks run higher. Probes are fixed-line boxes, so mobile data and Wi-Fi would add to every figure. Coverage is uneven: several states have no probe, and most places have one or two.

Check it. The Globalping results stay viewable for a few weeks after a run: Mumbai, Hyderabad, Singapore, Frankfurt, London, Virginia, California. Every probe's result is in the downloads below, permanently.

05Use the data

Free to cite.

Published under CC BY 4.0: use the figures anywhere, commercially included, with a link to this page.

Download CSV Download JSON

Cite as

FasterHost, "India server latency report", 25 September 2026, https://www.fasterhost.in/india-latency-report.php

Link to it

<a href="https://www.fasterhost.in/india-latency-report.php">India server latency report (FasterHost, 2026)</a>

06Questions

What people ask about it.

It works, but it is measurably slower. In this measurement a server in Singapore was a median 73 ms from Indian connections, against 35 ms for Mumbai, and every uncached request pays that difference more than once. The gap depends on where the visitor is; the city table shows it place by place.
Lower is always better, because a browser pays the round trip several times before a page starts to arrive. Hosting in India kept it to a median of 35 ms here. Singapore roughly doubled that, Europe multiplied it by about 4 and the United States by about 7.
Not on its own. Google works out which country a site is for from signals such as the domain, the language and hreflang, not from where the server sits. Location matters through speed: a distant server delays the first byte of every page, which delays Largest Contentful Paint, one of the Core Web Vitals.
Yes. The figures, charts and downloads are published under the Creative Commons Attribution 4.0 licence: use them anywhere, commercially included, with a link to this page as the source.

→Get started

Host where your visitors are.

cPanel, VPS and dedicated servers in Mumbai and Delhi NCR, with free managed migration and a 7-day money-back guarantee.

See pricing Talk to an engineer

Median support reply: 1m 48s