a2-highgpu-8g
$34.6768/hr
~$25314.03/month on-demand in us-south1
The a2-highgpu-8g is an Accelerator optimized instance with 96 vCPUs and 680 GiB of memory, 8x NVIDIA nvidia-tesla-a100 with up to 32 Gbps bandwidth, starting at $34.677 per hour or $29.209 per hour with spot instances.
Last updated: Jan 16, 2026, 12:03 PM UTC
Pricing Options
CUD
Committed Use Discounts
1 Year
$32.0937/hr
~$23428.38/mo
Save 7%
3 Year
$30.1395/hr
~$22001.87/mo
Save 13%
Spot Pricing
Up to 90% savings - flexible workloads
Spot Price
$29.2089/hr
~$21322.49/mo
Save 16%
a2-highgpu family
Switch between sizes in us-south1
| Size | vCPU | Mem |
|---|---|---|
| a2-highgpu-1g | 12 | 85 |
| a2-highgpu-2g | 24 | 170 |
| a2-highgpu-4g | 48 | 340 |
| a2-highgpu-8g | 96 | 680 |
Instance Specifications
Detailed hardware and capabilities for a2-highgpu-8g
Processor
Architecturex86_64
vCPUs96
Memory680 GiB
Memory per vCPU7.08 GiB
Network
PerformanceUp to 32 Gbps
Bandwidth32 Gbps
Max Interfaces8
Enhanced Networking
Storage
Persistent Disk Only
GPU
Count8
ManufacturerNVIDIA
Modelnvidia-tesla-a100
Memory40 GiB
Total GPU Memory320 GiB
Capabilities
Familya2-highgpu
GenerationCurrent
Burstable
Regional Pricing Comparison
Compare a2-highgpu-8g (Linux) pricing across all regions - sorted by cost
Current region (us-south1)Other regions
💡 Click any bar to open that region in a new tab