g7e.48xlarge

Amazon EC2 · G7e family

The g7e.48xlarge is a gpu & accelerated Amazon EC2 instance type with 192 vCPUs and 2 TiB of memory, 8 × NVIDIA RTX PRO 6000 Blackwell, running on Intel Xeon Scalable (Emerald Rapids) with 1600 Gigabit networking, starting at $33.14 per hour on demand.

gpu & acceleratedx86-64intelgpu
vCPUs
192

dedicated

memory
2 TiB

10.67 GiB per vCPU

gpu
8× RTX PRO 6000 Blackwell

768 GiB total

on-demand
$33.14

$24,195/mo · us-east-1

spot
–

not offered

regions
13

across 3 areas

g7e.48xlarge price by region

13 regions · hourly list prices in USD · On-demand · Spot · 1-yr Reserved · 3-yr Reserved

US East (N. Virginia)us-east-1cheapestreferenceNorth America$33.14
US East (Ohio)us-east-2cheapestNorth America$33.14
US West (Oregon)us-west-2cheapestNorth America$33.14
EU (Spain)eu-south-2Europe$34.93
Asia Pacific (Seoul)ap-northeast-2Asia$40.75
EU (Stockholm)eu-north-1Europe$47.81
Asia Pacific (Tokyo)ap-northeast-1Asia$48.07
Asia Pacific (Mumbai)ap-south-1Asia$54.14
EU (Frankfurt)eu-central-1Europe$56.36
Asia Pacific (Malaysia)ap-southeast-5Asia$56.78
Asia Pacific (Thailand)ap-southeast-7Asia$56.78
EU (London)eu-west-2Europe$57.61
Asia Pacific (Jakarta)ap-southeast-3Asia$63.10

Specifications

compute

vCPUs
192
Memory
2 TiB
Memory per vCPU
10.67 GiB
Processor
Intel Xeon Scalable (Emerald Rapids)
CPU vendor
Intel
Architecture
x86-64
Clock speed
3.2 GHz
Burstable CPU
No

storage & network

Local storage
15.2 TB NVME (4 disks)
Storage detail
4 x 3800 NVMe SSD
Network
1600 Gigabit
Peak bandwidth
1,600 Gbps

accelerators

Count
8
Memory (total)
768 GiB
Architecture
Blackwell
Vendor
NVIDIA

aws details

API name
g7e.48xlarge
Family
G7e
Category
GPU & accelerated
Generation
Current
Regions
13 (Asia, Europe, North America)
Cheapest region
US East (N. Virginia) · $33.14/hr
Dedicated EBS throughput
100000 Mbps
Enhanced networking
Yes
Normalization size factor
9.86
Processor features
Intel AMX
Spot interruption rate (us-east-1)
>20%
Spot savings vs on-demand (us-east-1)
68%

g7e.48xlarge equivalents on other clouds

closest match by vCPU, memory, CPU architecture, accelerator and class · Linux on-demand in each provider's reference region

compare side by side→
ProviderInstancevCPUMemoryProcessorOn-demandvs g7e.48xlarge
AWSg7e.48xlarge1922 TiB8 × NVIDIA RTX PRO 6000 Blackwell$33.14–
Google Clouda3-highgpu-8g208 vCPU (1.08×) · 1872 GiB (1.1× less) · NVIDIA H1002081,872 GiB8 × NVIDIA H100$88.49+167%
a3-megagpu-8g208 vCPU (1.08×) · 1872 GiB (1.1× less) · NVIDIA H1002081,872 GiB8 × NVIDIA H100$92.94+180%

G7e family sizes

all about G7e→
SizevCPUMemoryAcceleratorLocal storageOn-demandMonthly
g7e.2xlarge864 GiB1 × NVIDIA RTX PRO 6000 Blackwell1.9 TB NVME$3.363$2,455
g7e.4xlarge16128 GiB1 × NVIDIA RTX PRO 6000 Blackwell1.9 TB NVME$3.998$2,919
g7e.8xlarge32256 GiB1 × NVIDIA RTX PRO 6000 Blackwell1.9 TB NVME$5.268$3,846
g7e.12xlarge48512 GiB2 × NVIDIA RTX PRO 6000 Blackwell3.8 TB NVME$8.286$6,049
g7e.24xlarge961 TiB4 × NVIDIA RTX PRO 6000 Blackwell7.6 TB NVME (2 disks)$16.57$12,098
g7e.48xlarge1922 TiB8 × NVIDIA RTX PRO 6000 Blackwell15.2 TB NVME (4 disks)$33.14$24,195

g7e.48xlarge FAQ

How much does g7e.48xlarge cost?

g7e.48xlarge costs $33.14 per hour on demand with Linux in US East (N. Virginia) (us-east-1), about $24,195 per month. Across 13 regions the Linux on-demand price ranges from $33.14 in US East (N. Virginia) to $63.10 in Asia Pacific (Jakarta). With Windows it costs $41.98 per hour.

How many vCPUs and how much memory does g7e.48xlarge have?

g7e.48xlarge has 192 vCPUs and 2 TiB of memory (10.67 GiB per vCPU), running on Intel Xeon Scalable (Emerald Rapids). It is a gpu & accelerated instance type in the G7e family.

What GPU does g7e.48xlarge have?

g7e.48xlarge has 8 × NVIDIA RTX PRO 6000 Blackwell with 768 GiB of accelerator memory in total.

What are the equivalents of g7e.48xlarge on other clouds?

The closest matches by vCPU, memory, CPU architecture, accelerator and instance class are on Google Cloud, a3-highgpu-8g (208 vCPU, 1,872 GiB, $88.49/hour in Iowa).