Cloud gaming on the train — what actually works

You're on a 90-minute commute. You have a tablet. The train has WiFi or you have a 5G hotspot. You play Starfield or Cyberpunk or Forza Horizon on your commute via cloud streaming. The 90 minutes are filled with AAA gaming instead of doomscrolling.

This is one of the most-pitched cloud gaming use cases. It also turns out to be a use case where the country and the carrier matter dramatically.

Japan: Shinkansen on-board WiFi has been steadily improving and JR-East/JR-Central's network can sustain cloud gaming at 720p with acceptable latency for the duration of a Tokyo-Osaka run. Carrier 5G (Docomo, KDDI, SoftBank) is excellent and supports cloud gaming on-train.

South Korea: KTX (Korea's high-speed rail) on-board WiFi is exceptional. SK Telecom and KT both have 5G coverage along the major rail lines that supports cloud gaming reliably.

Switzerland and Germany: ICE and SBB trains have functional on-board WiFi with carrier-network fallback. Works for cloud gaming at 720p but variable.

Singapore MRT: excellent coverage, cloud gaming works for the entire commute. But MRT trips are short and the use case is closer to '20 minutes between meetings' than 'long commute'.

US commuter rail and Amtrak: WiFi unreliable, carrier 5G coverage spotty (especially in tunnels and along non-Acela routes). Cloud gaming on US transit is mostly aspirational.

UK trains: famously bad on-board WiFi, mixed carrier 5G performance. Cloud gaming attempts produce frequent disconnects.

Most subway/metro systems globally: tunnels block carrier signal, on-platform WiFi exists in some metros but cloud gaming requires sustained connection during travel that's not delivered.

Bus transit: variable everywhere. Carrier 5G handoff between cell towers during a bus ride is harsh on cloud gaming streams. Generally not workable.

Cloud gaming on transit isn't just about average bandwidth. Latency variance during cell handover is the killer. Each time the train passes between 5G cells, the connection re-establishes — 100-300 ms of dropped frames during the handover.

Cloud streaming protocols can absorb some packet loss but the visual stutter during cell handover is immediately noticeable. After 3-4 handovers in 10 minutes, the play experience becomes frustrating.

Japanese and Korean rail networks handle this best because the carrier coordination with the rail operators reduces handover frequency. Other markets haven't invested in this coordination.

Pre-downloaded content for the journey. If your game has an offline mode (most AAA single-player do), download a session's worth of progress before the trip and play locally. Doesn't help cloud-only catalogues but works for owned-game scenarios.

Latency-tolerant genres. Roguelikes (Slay the Spire, Balatro) work over high-variance connections because the gameplay doesn't depend on consistent millisecond response. Turn-based strategy similar. Action games that need quick reflexes do not work.

Shorter sessions accepted as the framing. Don't expect a 90-minute uninterrupted cloud gaming session on transit. Expect a series of 10-20 minute sessions punctuated by 1-3 minute disconnect windows. The mindset matters.

Cloud gaming on transit is a real use case for users in Japan, Korea, and parts of Western Europe with strong rail networks. It's mostly aspirational elsewhere.

If you're in the US, the UK, or most emerging markets, cloud gaming on transit will be a frustrating experience that delivers the marketing's promise sometimes and fails frequently. Plan for that.

The transit cloud gaming use case will improve over time as 5G and rail-network WiFi continue to roll out. The improvement is not on a 12-month horizon; it's on a 3-5 year horizon. Don't make a cloud gaming subscription decision based on transit-readiness today.