All products 
VPS in Tokyo: how Japan's main hub works and what to check in a plan
Tokyo is Japan's network capital and one of the main hubs of the Asia-Pacific region. Cables from the US, Korea, Taiwan, Hong Kong and Singapore converge here, the largest Japanese internet exchanges operate here, and almost every global cloud has a site here. That is why "Tokyo" appears in the location lists of Japanese hosts as well as American and European providers that need a single point in Asia.
Where "Tokyo" servers actually sit
The city's network core is the Otemachi business district and the neighbouring blocks in the centre: carrier nodes and exchange points are concentrated there. But building large halls for thousands of servers in the middle of a megacity is expensive, so most capacity has moved to the suburbs: Chiba, Saitama and Kanagawa prefectures. A cloud "Tokyo region" usually consists of several sites spread across the metro area. Submarine cables come ashore on the Chiba and Ibaraki coasts, not in the city itself.
For latency this hardly matters: within the metro area the difference is a few milliseconds. What matters is something else: in the catalogue the city comes from the host's list of sites, and for some plans "Tokyo" means the Kanto region as a whole. Before ordering, check:
- The data centre name in the plan description or on the host's website. If it isn't there, ask support.
- A test IP or looking glass in Tokyo, not a shared one for all Asian locations.
- The geolocation of the assigned address. Small hosts' addresses are often registered in the US or Europe, and services that detect the country by IP don't see Japan. If a Japanese IP matters for your task, confirm it before paying.
- The return route. Cheap transit sometimes leaves Japan via Los Angeles, so packets to Europe cross two oceans.
Routes and ping from Europe and the US
From the US west coast Tokyo is about 100 ms away, from the east coast usually 150–180 ms. From Western Europe it is typically 220–280 ms, and most of that spread comes from the route, not the distance: traffic goes either across the US and the Pacific or via the Middle East and the Indian Ocean. Which one you get depends on your provider and on the host's transit contracts, so two servers in the same Tokyo building can differ by tens of milliseconds for you. Look at a traceroute to the test IP: if a European connection goes through American nodes, the path is long.
Within the region Tokyo is very convenient: about 30 ms to Seoul, 30–50 ms to Taipei and Hong Kong, about 70 ms to Singapore. If your service targets both Asia and America, compare with a site in Los Angeles: it is closer to an American audience but further from a Japanese one.
Tokyo or another site in Japan
A general overview of the country (laws, payment, market specifics) is on the VPS in Japan page. Here we only cover the choice of city. Tokyo is the pick when you need the best connectivity to the US and the rest of Asia, the widest choice of configurations and a Japanese audience nationwide. Osaka makes sense as a second site for redundancy: if the Tokyo metro area loses connectivity or power, the project keeps running. Tokyo alone is enough for a single website or bot; for a service where downtime costs money, use two cities.
Who a Tokyo server suits
- Projects with Japanese users: stores, services, mobile backends where every 100 ms of latency is noticeable.
- Game servers for players in Japan, Korea and Taiwan; CPU and bandwidth requirements are covered on the VPS for a game server page.
- Testing from Asia: checking site speed, regional ads and app store results from a Japanese address.
- An Asian node paired with Europe or the US: if your users are spread worldwide, Tokyo covers East Asia, while for a focus on Southeast Asia Singapore is a better fit.
In the evening Japanese time, when home traffic rises, plans with cheap transit can see latency spikes. If your service is sensitive to stability, look for plans that name their carriers and don't judge by a single daytime measurement.
One more detail is the time zone. Tokyo data centres carry out planned maintenance at night Japanese time, which is the late afternoon or evening in Europe and around midday on the US east coast: maintenance windows announced by the host may fall into peak hours for your Western audience. If the server serves both Asia and Europe or the US, check the maintenance schedule in advance and warn your users.