Fireworks adds GLOBAL multi-region deployments under one endpoint

Fireworks AI introduced a GLOBAL option that lets one inference deployment run across geographies behind a single endpoint. The scheduler places workloads across eligible capacity while respecting hardware, quota, reliability, and data residency constraints.

In a seven-day observational study, deployments spread across two or more serving clusters had a 99.992% request success rate, compared with 99.269% for single-region deployments.