GPU Cloud Provider · Mountain View, California, USA
Google Cloud
Google Cloud provides a robust and secure global infrastructure designed to deliver high performance and reliability for various applications, including VM deployment, AI, ML, and HPC workloads. It benefits from a private global network, environmentally sustainable practices, and compliance with stringent security standards.
GPUs
42
Founded
First offered in 2008
Countries
19
Data Centers
32
GPU Marketplace
NVIDIA H100 SXMOn-Demand
NVIDIA H100 SXMOn-Demand
NVIDIA H100 SXMOn-Demand
NVIDIA H100 SXMOn-Demand
NVIDIA H100 SXMOn-Demand
NVIDIA H100 SXMOn-Demand
NVIDIA H100 SXMOn-Demand
NVIDIA H100 SXMOn-Demand
NVIDIA H100 SXMOn-Demand
NVIDIA H100 SXMOn-Demand
NVIDIA H100 SXMOn-Demand
NVIDIA H100 SXMOn-Demand
Estimates only — rates are collected automatically from public provider pages and may be out of date. Prices vary by region, commitment term, and availability, and typically exclude storage, egress, and tax. Confirm current pricing with the provider before purchasing. Last collected .
Company Profile
Provider TypeCompute Engine
FoundedFirst offered in 2008
HeadquartersMountain View, California, USA
Legal EntityGoogle Cloud
Infrastructure
GPU FleetNVIDIA RTX PRO 6000, GB300, GB200, B200, H200, H100, L4, P100, P4, T4, V100, and A100 GPUs
Network FabricPrivate global backbone
ConnectivityUp to 3,600 Gbps for specific machine types
StorageLocal SSD, Persistent Disk
AvailabilityGA
machine learning, scientific computing, and generative AI
Compute & Deployment
On-DemandAll with the per-second billing, so you only pay only for what you need while you are using it.
Reserved InstancesFor GPUs, future reservations in calendar mode and Flex-start VMs are available in all zones where the GPU machine types are available.
VM-BasedWhen you create Compute Engine (VM) instances or other resources on Google Cloud, you can attach accelerators to them.
Container-BasedThis document focuses on accelerator availability for Compute Engine, Google Kubernetes Engine (GKE), and AI Hypercomputer.
KubernetesThis document focuses on accelerator availability for Compute Engine, Google Kubernetes Engine (GKE), and AI Hypercomputer.
GPU Hardware
Latest GenNVIDIA RTX PRO 6000, GB300, GB200, B200, H200, H100, L4, P100, P4, T4, V100, and A100
Multi-GPU Nodesup to 8 GPUs per instance
Max GPUs/Nodeup to 8 GPUs per instance
Pricing Model
Subscriptionper-second billing
Public PricingFor information about GPU pricing for the different GPU types and regions that are available on Compute Engine, refer to the GPU pricing document.
Pay-as-you-goAll with the per-second billing, so you only pay only for what you need while you are using it.
Credit SystemStart building on Google Cloud with $300 in free credits and 20+ always free products.
Performance & Scaling
Developer Experience
OnboardingStart building on Google Cloud with $300 in free credits and 20+ always free products.
FrameworksTensorFlow, PyTorch
TemplatesAvailable pre-built solutions
DocumentationDocumentation includes 'GPUs on Compute Engine' and multiple tutorials such as 'Adding or removing GPUs on Compute Engine', 'Installing GPUs drivers', 'GPUs on Google Kubernetes Engine', 'Using GPUs for training models in the cloud', 'Attaching GPUs to Managed Service for Apache Spark clusters', and 'Using GPUs with Dataflow'.
API FeaturesCLI, SDK, REST API
Security & Compliance
SecurityRegular security assessments,Compliance certifications from third-party auditors
ComplianceStringent security standards adherence reviewed by third-party auditors
Data Center Locations
Coverage
CountriesTaiwan, Hong Kong, Japan, South Korea, India, Indonesia, Singapore, Australia, Poland, Finland, Belgium, England, Germany, Netherlands, Switzerland, Italy, Saudi Arabia, Israel, Brazil
CitiesChanghua County, Hong Kong, Tokyo, Seoul, Mumbai, Delhi, Jurong West, Jakarta, Sydney, Warsaw, Hamina, St. Ghislain, London, Frankfurt, Eemshaven, De Kooy, Zurich, Milan, Dammam, Tel Aviv, Montréal, Toronto, Osasco, Council Bluffs, Moncks Corner, Ashburn, Columbus, Dallas, The Dalles, Los Angeles, Salt Lake City, Las Vegas
APAC North America South America Europe Middle East Africa AI zone
Compliance Regions
India Regionasia-south1,asia-south2
Datacenter Locations
Key Strengths
Wide selection of GPU types
flexible pricing and machine customizations
balance processor/memory/disk with up to 8 GPUs per instance
per-second billing
integration with Google Cloud storage, networking and data analytics.
Known Limitations
GPU model availability varies by region/zone
some machine types (for example A3 Mega
A3 High, a2-megagpu-16g) have limited capacity and require contacting your account team
some fractional G4 machine types are only available in specific zones
Local SSDs have zone-specific exceptions with some GPU machine types.
Additional Information
Support Options
["24/7 support via phone, email, and online resources"]
Community
Newsletter and blog presence (product updates and blog posts).
Core Proposition
A wide selection of GPUs to match a range of performance and price points
Last updated March 2026. Information subject to change.