Fluidstack

Managed GPU clusters for AI labs, contracted not self-serve

HQ: USA
Founded: 2017
Locations: 4

Visit Fluidstack Compare with another provider

What's good about Fluidstack

Managed Slurm and Kubernetes clusters on bare metal
Active and passive cluster health checks with observability
ClusterMAX Gold rating from SemiAnalysis
Custom data centers built for a single customer's workload

Pricing & billing

Fluidstack publishes no public rates. Capacity is contracted directly and quoted per deal, and there is no self-serve hourly option or online signup.

Lyceum

NVIDIA H200 from $2.20/hr, B200 from $3.40/hr, B300 from $3.80/hr

Per-second billing with no minimum commitment or vendor lock-in

Scale to 8,000 GPUs over InfiniBand for distributed AI training

View Pricing

About Fluidstack

Fluidstack is a New York headquartered AI cloud provider that builds and operates data centers for frontier model training. Founded in 2017, it sells GPU capacity as managed Slurm and Kubernetes clusters on bare metal, contracted directly with AI labs, governments and enterprises rather than sold self-serve. Anthropic selected the company in November 2025 to deliver custom data centers in New York and Texas, with sites coming online through 2026.

Services

Fluidstack offers 1 of the 27 services we track.

Service	Provider page
GPU-powered Servers	On Fluidstack

Locations

Based on our records, Fluidstack operates in 4 locations across 4 countries.

Country	Location	Region code
Canada	Canada	CANADA
Netherlands	Netherlands	NETHERLANDS
Norway	Norway	NORWAY
Sweden	Sweden	SWEDEN

Alternatives to Fluidstack

Compare Fluidstack against other cloud providers.

More comparisons

Our data for Fluidstack was last updated on Sept. 18, 2026.