GPU Cloud Provider · Redwood Shores, California, USA
Oracle Cloud Infrastructure
Oracle Cloud Infrastructure (OCI) provides advanced AI infrastructure designed for a wide range of AI workloads, including model training and inference. OCI supports scale-out capabilities up to zettascale with bare metal instances, equipped with the latest NVIDIA GPUs and high-bandwidth, low-latency RDMA networking.
Founded
Oracle Corporation founded in 1977, with Oracle Cloud developed later.
Countries
55
Company Profile
Provider TypeOracle Cloud Infrastructure (OCI)
FoundedOracle Corporation founded in 1977, with Oracle Cloud developed later.
HeadquartersRedwood Shores, California, USA
Legal EntityOracle Cloud Infrastructure (OCI)
Parent CompanyOracle
Infrastructure
GPU FleetMultiple GPU shapes including BM.GPU.RTXPro.8 (8 x NVIDIA RTX PRO 6000 96GB), BM.GPU.B200.8 (8x Nvidia B200 180GB), BM.GPU.GB200.41 (4x Nvidia B200 189GB), BM.GPU.GB300.42 (4x Nvidia B300 279GB), BM.GPU.B300.8 (8x Nvidia B300 263GB), BM.GPU.H200.8 (8x Nvidia H200 141GB Tensor Core), BM.GPU.H100.8 (8x NVIDIA H100 80GB Tensor Core), BM.GPU.MI300X.8 (8x AMD MI300X 192 GB Matrix Core), BM.GPU.MI355X.8 (8x AMD MI355X 288GB Matrix Core), BM.GPU.A100-v2.8 (8x NVIDIA A100 80GB Tensor Core), BM.GPU.L40S.4 (4x NVIDIA L40S 48GB), BM.GPU4.8 (8x NVIDIA A100 40GB Tensor Core), VM.GPU.A10.1 (1x NVIDIA A10), VM.GPU.A10.2 (2x NVIDIA A10), BM.GPU.A10.4 (4x NVIDIA A10), VM.GPU3.1 (1x NVIDIA V100), VM.GPU3.2 (2x NVIDIA V100), VM.GPU3.4 (4x NVIDIA V100), BM.GPU3.8 (8x NVIDIA V100), VM.GPU2.1 (1x NVIDIA P100), BM.GPU2.2 (2x NVIDIA P100)
Network FabricRDMA over Converged Ethernet (RoCE v2)
ConnectivityUp to 3,200 Gbps of cluster network bandwidth, up to 400 Gbps front-end network bandwidth
StorageNVMe SSD, Oracle-managed file storage with Lustre, Block storage, Object storage
Bare MetalInstances are available as virtual machines and bare metal
AvailabilityGA
AI and Machine LearningAnalytics and BIComputeContainers and FunctionsCustomer Success ServicesData LakeDeveloper ServicesHybrid CloudIntegrationNative VMwareNetworkingObservability and ManagementOpen Source DatabasesOracle DatabasesPayment Cloud ServiceSecurityStorageOthers
Compute & Deployment
On-DemandYes (foundational models can be consumed on demand)
Spot / InterruptibleYes (preemptible instances available)
Reserved InstancesYes (Unused Capacity reservations available)
Bare MetalYes (bare metal instances available)
VM-BasedYes (virtual machine instances available)
Container-BasedYes (container instances available)
KubernetesYes (Cloud Infrastructure Kubernetes Engine)
GPU Hardware
Latest Gen| Oracle Cloud Infrastructure NVIDIA AI Enterprise - H100 | | GPU per hour |, | Oracle Cloud Infrastructure NVIDIA AI Enterprise - L40S | | GPU per hour |, | Oracle Cloud Infrastructure NVIDIA AI Enterprise - A10 | | GPU per hour |, | Oracle Cloud Infrastructure NVIDIA AI Enterprise - A100 80 | | GPU per hour |, | Oracle Cloud Infrastructure NVIDIA AI Enterprise - GB200 | | GPU per hour |, | Oracle Cloud Infrastructure NVIDIA AI Enterprise - B200 | | GPU per hour |, | Oracle Cloud Infrastructure NVIDIA AI Enterprise - H200 | | GPU per hour |, | Oracle Cloud Infrastructure NVIDIA AI Enterprise - A100 40 | | GPU per hour |
Legacy SupportPrevious generation instances are in use by some current customers but are no longer available for new deployments.
Multi-GPU NodesYes (examples include multi-GPU shapes such as BM.GPU.H200.8 with 8 GPUs)
Max GPUs/Node8
InfiniBand4x400 Gb/sec InfiniBand (on BM.GPU.GB200.41)
Pricing Model
Per HourOCPU per hour
Per MinuteConnection time minute
SubscriptionUniversal Credits Consumption (% applied to consumption, with minimum)
Reserved DiscountUnused Capacity reservations are priced at 85% of the price of regular instances. This discount is applied in addition to other discounts, such as UCM negotiated rates.
Spot DiscountPreemptible instances are priced at 50% of the price of regular instances. This discount is applied in addition to other discounts, such as UCM negotiated rates.
Public Pricinghttps://oracle.com/cloud/compute/pricing/
Hidden FeesWindows Server on-demand license cost is an add-on to the underlying compute instance price. You will pay for the compute instance cost and Windows license cost separately.
Credit SystemUniversal Credits Consumption
Performance & Scaling
Multi-Node TrainingLarge scale-out AI training, data analytics, and HPC
InfiniBand4x400 Gb/sec InfiniBand
Developer Experience
OnboardingTry Oracle Cloud Free Tier; paid tenancy free allotment for Ampere A1: first 3K OCPU-hrs and first 18K GB-hrs per month
FrameworksOCI Data Science
JupyterData Science Notebook Sessions
Model MarketplaceOCI Generative AI with named providers (Cohere, Meta, xAI, Google, OpenAI); foundational models consumable on demand; ability to host private replicas and fine-tune on dedicated AI clusters
DocumentationReferences to Compute FAQ and product documentation; guidance to 'Get additional details' on OCPU vs vCPU; mentions product documentation for supported shapes
API FeaturesCLI, SDK, REST API
Security & Compliance
SecurityComplies with industry-standard security and compliance protocols suitable for government and dedicated clouds.
ComplianceData residency, security, and compliance capabilities are supported, particularly for government and sovereign cloud services.
Lists integrations/offerings from CohereMeta, xAIGoogleOpenAImentions Oracle Linux Premier Support
Data Center Locations
Coverage
CountriesAUD - Australian Dollar ($), CAD - Canadian Dollar ($), EUR - Euro (€), GBP - British Pound (£), USD - US Dollar ($), AED - United Arab Emirates (د.إ), ALL - Albanian (Lek), ARS - Argentine Peso ($), BAM - Bosnia-Herzegovina Convertible Mark (KM), BDT - Bangladesh Taka (৳), BHD - Bahraini Dinar (.د.ب), BRL - Brazilian Real (R$), CHF - Swiss Franc (CHF), CLP - Chilean Peso ($), CNY - Chinese renminbi (¥), COP - Colombian peso ($), CRC - Costa Rican colón (₡), CZK - Czech Koruna (Kč), DKK - Danish Krone (kr), EGP - Egyptian pound (£), HKD - Hong Kong Dollar (HK$), HRK - Croatian Kuna (kn), HUF - Hungarian Forint (Ft), IDR - Indonesian Rupiah (Rp), ILS - Israeli New Sheqel (₪), INR - Indian Rupee (₹), ISK - Icelandic Króna (kr), JPY - Japanese Yen (¥), KES - Kenyan Shilling (Ksh), KRW - South Korean Won (₩), KWD - Kuwaiti Dinar (د.ك), KZT - Kazakhstani tenge (₸), MOP - Macanese Pataca (P), MVR - Maldivian Rufiyaa (Rf), MXN - Mexican Peso ($), MYR - Malaysian Ringgit (RM), NOK - Norwegian Krone (kr), NZD - New Zealand Dollar (NZ$), OMR - Omani rial (ر.ع.), PEN - Peruvian Nuevo Sol, PHP - Philippine Peso (₱), PKR - Pakistani rupee (Ɍs), PLN - Polish Zloty (zł), QAR - Qatari Riyal (ر.ق), RON - Romanian Leu (lei), RSD - Serbian Dinar (Дин.), RUB - Russian Ruble (₽), SAR - Saudi Riyal (ر.س), SEK - Swedish Krona (kr), SGD - Singapore Dollar ($), THB - Thai Baht (฿), TRY - Turkish Lira (TL), TWD - New Taiwan Dollar (NT$), VND - Vietnamese Dong (₫), ZAR - South African Rand (R)
Compliance Regions
Datacenter Locations
Key Strengths
Customer-friendly pricing to save money
innovative capabilities to avoid overpayment and billing surprises
support for on-demand foundational models and hosting private replicas on dedicated AI clusters
Known Limitations
Many pricing table unit price cells are blank on this page
previous-generation instances are not available for new deployments
some HPC shapes require contacting sales
Additional Information
Support Options
Dedicated support, online resources
Core Proposition
customer-friendly approach to saving money in the cloud; innovative capabilities to eliminate needless overpayment and billing surprises
Last updated March 2026. Information subject to change.