Customer story
Arcee AI selects Fuzzball for portable workload orchestration at frontier scale
Arcee coordinates compute-intensive model-development workloads across cloud and on-premises environments while its team keeps control of its models, data, and training and inference stack.
“We have built our own training and inference stack because control over the full model lifecycle is fundamental to how Arcee operates.”
The scope, the environments, and where the line sits
What Fuzzball orchestrates for Arcee, and what Arcee keeps under its own control.
- Customer
- Arcee AI, the U.S. model lab behind the Trinity family of models
- Scope
- Workload orchestration for the development of Arcee’s next flagship large language model
- What Fuzzball does
- Coordinates and places selected compute-intensive workloads across heterogeneous infrastructure
- What Arcee retains control of
- Training and inference software, models, data, and deployment architecture
- Environments
- Cloud, GPU-provider, and on-premises
The story
Arcee AI builds open-weight foundation models end to end, including Trinity Large, a 400-billion-parameter sparse mixture-of-experts model.
Arcee selected Fuzzball to coordinate and place selected compute-intensive workloads across cloud, GPU-provider, and on-premises environments, so its model-development systems stay independent of the underlying compute provider. Arcee’s engineers continue to run the training and inference stack they built. Fuzzball gives them one consistent orchestration layer beneath it, and Arcee retains control of its models, data, and deployment architecture.
“Fuzzball gives us an additional orchestration layer for coordinating workloads across heterogeneous compute environments without replacing the systems our team has built.”
How Fuzzball fits
One orchestration layer underneath the stack Arcee already built, not a replacement for it.
One orchestration layer across three environments.
The same layer coordinates workloads across cloud, GPU-provider, and on-premises infrastructure.
Independence from the compute provider.
Arcee’s model-development systems remain independent of the underlying compute provider.
A clear division of responsibility.
Arcee’s team decides what runs. Fuzzball helps the team manage where and how selected jobs are executed.
Go deeper
Read the announcement
The full press release on Arcee’s selection of Fuzzball.
See how Fuzzball works
Architecture, capabilities, and deployment options for portable workload orchestration.
Sovereign AI at CIQ
How CIQ approaches AI infrastructure organizations own and operate.
Fuzzball sovereign AI technical brief
The full technical picture of running sovereign AI workloads on infrastructure you control.
Where this goes next
Arcee and CIQ will collaborate to help enterprises, AI labs, and infrastructure providers deploy models on portable infrastructure they control. Arcee brings model development, adaptation, and the training and inference systems it builds and operates itself. CIQ brings Enterprise Linux, cluster management, and workload-orchestration technologies that can help those systems run across a range of compute environments.
About Arcee AI
Arcee AI is a U.S. open-model lab that develops open-weight foundation models end to end, including data preparation, architecture, pretraining, post-training, evaluation, and deployment. Its Trinity family ranges from compact models designed for local use to Trinity Large, a 400-billion-parameter sparse mixture-of-experts model. Arcee builds models that developers, enterprises, and public institutions can run and adapt on infrastructure they control.
The Race to Build an American Alternative to Cheap AI From China
August 2026
The Top Open AI Models Are Chinese. Arcee AI Thinks That’s A Problem.
Anna Tong, February 2, 2026
Run your models on infrastructure you control
Talk to a CIQ engineer about orchestration for AI and HPC workloads across cloud and on-premises environments.