Regions ↗
noOriginal Documentation
Documentation Index#
Fetch the complete documentation index at: https://docs.fireworks.ai/llms.txt Use this file to discover all available pages before exploring further.
Fireworks runs a global fleet of hardware on which you can deploy your models.
Fireworks runs a global fleet so you can deploy models close to users, meet data-residency needs, and scale across clouds. This page covers multi-region (default behavior and quota groupings), single-region availability and hardware, how to use and change regions, and quotas.
Multi-region (recommended)#
By default, deployments are multi-region: Fireworks can move and spread them across regions as needed. Multi-regions (GLOBAL, US, EUROPE, APAC) are high-level groupings of single regions. Your deployment may run in any single region(s) within that multi-region.
<span class=“card-start” data-card-raw=“title=“Flexible, elastic scaling” icon=“gauge”"> Utilizing multiple clouds and locations maximizes the odds that there’s capacity to scale.
<span class=“card-start” data-card-raw=“title=“Higher reliability” icon=“shield-check”"> Multi-region deployments enable resilience to localized outages, maintaining application availability as workloads scale across regions.
Supported multi-regions#
Supported multi-regions: GLOBAL, US, EUROPE, APAC.
Single region availability#
Single regions are concrete locations (e.g. US_IOWA_1, EU_FRANKFURT_1) where your deployment can run. We have the single regions listed below available; we recommend multi-region for most users because of its advantages (elastic scaling, higher reliability). If you have a specific need for a single region, contact Fireworks to request it. The table below shows which single regions are available and what hardware is offered in each.
| Region | Accelerator Type(s) | |
|---|---|---|
US_ARIZONA_1 | NVIDIA_H100_80GB | |
US_CALIFORNIA_1 | NVIDIA_H200_141GB | |
US_GEORGIA_2 | NVIDIA_B200_180GB | |
US_GEORGIA_3 | NVIDIA_H200_141GB | |
US_ILLINOIS_1 | NVIDIA_H100_80GB | |
US_ILLINOIS_2 | NVIDIA_A100_80GB | |
US_IOWA_1 | NVIDIA_H100_80GB | |
US_OHIO_1 | NVIDIA_B200_180GB | |
US_TEXAS_2 | NVIDIA_H100_80GB | |
US_UTAH_1 | NVIDIA_B200_180GB | |
US_VIRGINIA_1 | NVIDIA_H100_80GB, NVIDIA_H200_141GB | |
US_WASHINGTON_2 | NVIDIA_H100_80GB | |
US_WASHINGTON_3 | NVIDIA_B200_180GB | |
US_WASHINGTON_4 | NVIDIA_B200_180GB | |
EU_FRANKFURT_1 | NVIDIA_H100_80GB | |
EU_ICELAND_1 | NVIDIA_H200_141GB | |
EU_ICELAND_2 | NVIDIA_B200_180GB, NVIDIA_H200_141GB | |
AP_TOKYO_1 | NVIDIA_H100_80GB | |
AP_TOKYO_2 | NVIDIA_H200_141GB |
Using a region#
When creating a deployment, you can pass the --region flag to pin it to a single region:
firectl deployment create accounts/fireworks/models/llama-v3p1-8b-instruct \
--region GLOBALChanging regions#
Updating the single region for a deployment in-place is not supported. To move a deployment to a different single region, create a new deployment in the desired region, then delete the old deployment.
Quotas#
Quota is granted at the multi-region level for new users. By default, all users receive quota for GLOBAL multi-region. For specific single region quota, please contact Fireworks. To view your current quotas, run:
firectl quota listTo use single regions that are not generally available (see the table above), or to request additional multi-region quota, contact inquiries@fireworks.ai.