bring the workload, keep everything else
Open a new region without buying capacity for the peak. Bring a game, a model or a container; we run it near your users, per session or by the hour. You keep the surface, the users and the money.
every session has a person waiting on it
A game frame, a container's response or a spoken reply: each one runs on a GPU near the person waiting for it.
three product lines, one network
Games and live AI stream per session. Your own container runs by the hour. Same machines, same placement; what differs is what you hand us.
01
YOM Games · streaming, per session
a game
Playable on your own surface or behind your brand, on a GPU near the player.
publishers and platforms
the publisher route02
YOM AI · streaming, per session
a live conversation
Your own voice models, running in a session near the speaker.
voice teams
live voice sessions03
YOM Containers · per hour
a container
Your stack in its own container, placed near the user and billed by the hour while it runs.
platforms and server hosts
managed capacitystreaming by the session, containers by the hour
Streaming is billed per session of up to an hour, one fee however early it ends. A container holds its GPU for every hour it runs, so it is billed by the hour. No fleet to reserve, no multi-year term.
in a data centre
$1.23
a GPU hour, reserved for the peak
streaming, per session
$0.08
an hour of Standard play, one session (a session runs up to an hour)
classes from $0.04 (Light) to $0.24 (Ultra) a session
containers, per hour
$0.24
an hour of a Standard container, billed while it runs
classes from $0.12 (Light) to $0.72 (Ultra) an hour
A session is one player’s game or one live conversation. A container is billed for every hour it runs.
who is already asking
-
games
publishers and studios who want a playable game on every surface they own, on the games route
-
voice
voice teams that own their models and want them near the speaker
-
containers
streaming platforms with their own front end, and a low-latency server host that wants capacity next to its users
the ask
bring the workload
Tell us what it is, where your users are and how much of it there is. We come back with what the network serves today and what it costs.
talk to us about your workload