Neoclouds / Company profile
Fluidstack is a private AI infrastructure company building and operating large GPU compute capacity and the software layer used to run frontier-model training and inference on it. The company describes its platform as an AI cloud designed for “security, reliability, and scale” requirements typical of top AI labs, with managed orchestration options including Kubernetes (for training/inference) and Slurm (for frontier model training).
Company record
Forbes reported a previously unreported $1.5B round led by Jane Street and a valuation of over $18B, citing sources close to the company.
Fluidstack announced it raised a $830M Series A at a $7.5B valuation, stating the round was led by Situational Awareness.
Fluidstack announced a Midtown Manhattan headquarters relocation and stated it would create 300 permanent jobs plus over 800 construction jobs in New York State tied to a data center project plan referenced in the release.
Fluidstack announced a deal with Anthropic to build custom data centers in New York and Texas, stating the project would create approximately 800 permanent jobs and 2,400 construction jobs, with sites coming online throughout 2026.
Fluidstack announced it was awarded ClusterMAX Gold by SemiAnalysis, describing its ranking and responsiveness during an evaluation/benchmark process.
Forbes reported on a large AI infrastructure deal referenced as an “AI factory” expansion, framing Fluidstack’s positioning and market context.
Fluidstack announced a strategic GPU financing partnership with Macquarie to supply GPUs to premier AI research labs across Europe.
Fluidstack announced deployment and management of exascale clusters across Iceland and Europe with Borealis Data Center, Dell Technologies, and NVIDIA, including a stated figure for GPUs under management.
Still resolving: Newsroom · Official X
Fluidstack’s managed Kubernetes is described as an ultra-reliable bare metal container orchestration platform powering frontier model training and inference, with support for NVIDIA GPU Operator and Network Operator, plus passive and active health checks and comprehensive observability.
Fluidstack’s managed Slurm is described as an ultra-reliable bare metal batch orchestration platform powering frontier model training, including user management, Pyxis/Enroot, passive and active health checks, comprehensive monitoring, and topology-aware scheduling to minimize collective communication latency.
Lighthouse is Fluidstack’s monitoring/observability/reliability suite for Fluidstack infrastructure, including managed metrics and Grafana dashboards, passive health checks using NVIDIA DCGM background health checks, and additional reliability features such as active health checks (preview) and auto-remediation (preview).
Fluidstack publicly announced its $830M Series A led by Situational Awareness at a $7.5B valuation.
Situational AwarenessFluidstack announced establishment of a Midtown Manhattan office described as its global headquarters.
Fluidstack announced it was selected by Anthropic to deliver custom data centers in New York and Texas.
AnthropicFluidstack announced it was awarded ClusterMAX Gold by SemiAnalysis.
SemiAnalysisFluidstack announced deployment and management of exascale GPU clusters across Iceland and Europe in collaboration with NVIDIA, Borealis Data Center, and Dell Technologies.
NVIDIA · Borealis Data Center · Dell Technologies