GPU Cloud Provider · Unknown
Crusoe Cloud
Crusoe Cloud provides a robust and efficient GPU cloud platform optimized for AI and machine learning workloads. It features advanced automated management, robust fault detection and mitigation, and a highly scalable orchestration environment supporting frameworks like Kubernetes and Slurm. Crusoe Cloud leverages cutting-edge hardware such as NVIDIA Quantum-2 InfiniBand and offers a variety of storage options.
GPUs
2
Founded
Unknown
Countries
1
Data Centers
5
Team Size
201-1000
GPU Marketplace
NVIDIA H100 NVLOn-Demand
AMD Radeon RX 9070On-Demand
Company Profile
Company TypeScale-up
Provider TypeCloud Provider
Legal EntityCrusoe Energy Systems, Inc.
FundingSeries C/D
Total Raised$1.1B+
Team Size201-1000
Founders FundLux CapitalValor Equity PartnersPeter ThielBain Capital VenturesUpper90
Infrastructure
GPU FleetNVIDIA H100 SXM5 80GB, NVIDIA H100 SXM 80GB, NVIDIA A100 80GB, NVIDIA L40S 48GB, NVIDIA H200 SXM 141GB
Network FabricNVIDIA Quantum-2 InfiniBand
ConnectivityHigh-speed due to NVIDIA Quantum-2 InfiniBand
StorageEphemeral Disks, Persistent Disks, Shared Disks
Data Center TierTier 3 equivalent; proprietary and colocation facilities
Bare MetalYes, bare metal GPU instances available for high-performance AI workloads
AvailabilityGA
EnterpriseResearchAI/ML StartupsGovernment
Compute & Deployment
On-DemandYes
Spot / InterruptibleNo
Reserved InstancesYes (committed use contracts available for enterprise customers)
Bare MetalYes
VM-BasedYes
Container-BasedYes (Docker)
KubernetesYes (managed K8s)
Serverless GPUNo
Spin-Up Time2-5 minutes
TerraformYes (official provider)
GPU Hardware
Latest GenH100 SXM5
Legacy SupportA100 SXM4
Multi-GPU NodesYes (up to 8x per node)
Max GPUs/Node8
NVLinkYes (NVLink 4.0 on H100 SXM5 nodes)
InfiniBandYes (HDR 200Gbps)
PCIe vs SXMSXM only
HGX PlatformYes (HGX H100 8-GPU)
Pricing Model
Per HourYes (primary billing unit)
Per MinuteNo
SubscriptionYes (reserved instances with 1-year and 3-year terms)
Reserved DiscountUp to ~45% off with 1-year commitment
Spot DiscountNo spot pricing publicly disclosed
Public PricingPartial (some GPUs listed)
Hidden FeesNone disclosed
Pay-as-you-goYes
Credit SystemNo
Performance & Scaling
Multi-Node TrainingYes (multi-node distributed training supported with NCCL over high-speed interconnects)
Elastic ScalingManual only
Auto ScalingNo
InfiniBandYes (InfiniBand networking available on H100 and A100 SXM clusters)
NVSwitchYes (on SXM nodes)
Perf IsolationYes (dedicated bare metal instances)
Noisy NeighborYes (bare metal, no sharing)
Developer Experience
OnboardingSelf-service signup via web console; instances deployable in minutes; enterprise onboarding available with dedicated support
FrameworksNot explicitly stated but suggests support for popular AI frameworks
SDK LanguagesPython, Go
CLI ToolingFull CLI (crusoe CLI) for instance lifecycle management, SSH access, and resource configuration
JupyterVia SSH port forwarding or JupyterHub on provisioned instances
TemplatesPyTorch Training, LLM Fine-tuning, Triton Inference Server, NCCL Multi-node Workloads
DocumentationComprehensive docs with API reference, CLI guides, and deployment tutorials
API FeaturesPrometheus-compatible query API
Security & Compliance
SecurityEncrypted at rest data storage, but no explicit compliance certifications mentioned
ComplianceData encrypted at rest
Series C/D backed by Founders Fund and Lux CapitalStrategic partnership with NVIDIACustomers include frontier AI labs (xAI, Mistral)SOC 2 compliancePublished sustainability and carbon reduction metricsFeatured in major press (WSJ, Forbes, TechCrunch)
Data Center Locations
Coverage
CountriesUnited States
CitiesMidland TX, Abilene TX, Denver CO, Wyoming, North Dakota
North America
Compliance Regions
EU Data ResidencyNo EU presence
US Gov CloudNo
India RegionNo
Datacenter Locations
Key Strengths
Climate-aligned infrastructure using stranded/low-carbon energy
Competitive H100 and H200 pricing relative to hyperscalers
Strong focus on large-scale AI training workloads
Purpose-built for frontier AI with high-bandwidth networking
Backed by top-tier investors with long-term infrastructure commitments
Known Limitations
Primarily US-based regions; limited global footprint
No spot/preemptible GPU instances
Smaller ecosystem compared to AWS, GCP, Azure
Limited managed ML platform services (more IaaS-focused)
Less mature marketplace and tooling than hyperscalers
Additional Information
Support Options
["Dedicated support for production AI workloads","Advanced transparency into resource utilization and costs"]
Community
Limited public community; primarily enterprise-focused with dedicated Slack channels for enterprise customers and a public Discord for developers
Green Energy
Founded on stranded energy reduction mission; transitioning data centers to low-carbon and renewable energy sources; claims significant CO2e reduction vs. grid-powered competitors
Core Proposition
Sustainable GPU cloud powered by stranded and renewable energy sources, offering lower carbon footprint for AI and HPC workloads
Notable Customers
xAI
Mistral AI
Scale AI
Payment Methods
Credit CardWire TransferEnterprise Invoicing
Last updated March 2026. Information subject to change.