Crusoe, the first vertically integrated AI infrastructure provider, has announced two new AI managed services on its Crusoe Cloud platform
Crusoe Managed Inference – A fully managed, scalable AI inference service.
Crusoe AutoClusters – A fault-tolerant orchestration platform for AI training.
Both services, accelerated by NVIDIA, are being showcased at the NVIDIA GTC AI Conference.
“We’re eliminating the operational burdens that hinder AI innovation,” said Nadav Eiron, SVP of Cloud Engineering. “These new services allow developers to focus on AI breakthroughs, not infrastructure.”
Crusoe Managed Inference: AI Intelligence as a Service
Rapid AI Deployment – Run models instantly, without managing infrastructure.
Agentic AI Workflows – Seamlessly integrate AI responses into applications.
Intuitive UI – Test models and refine use cases through an easy-to-use interface.
“Crusoe Managed Inference enables developers to focus on intelligent applications instead of managing servers,” added Eiron.
Crusoe AutoClusters: Scalable AI Training Without Complexity
Effortless GPU Provisioning – Deploy NVIDIA Quantum-2 InfiniBand clusters with VAST Data-backed storage in one click.
Automated Fault Tolerance – Intelligent node replacement and job re-queueing.
Proactive Monitoring – Real-time diagnostics with NVIDIA DCGM integration.
Intelligent Job Scheduling – Fully managed Slurm & Kubernetes orchestration.
“We spun up 1,600 GPUs and ran jobs on Crusoe Cloud via Slurm—it just worked,” said Less Wright, PyTorch Partner Engineer at Meta.
Why This Matters
With AI adoption surging, Crusoe Cloud delivers:
Faster AI deployment with zero infrastructure hassle
Optimized high-performance computing with fault tolerance
Seamless AI model training & inference at scale
