← All instances

inf2.8xlarge

⚡ 1x GPU • 32.0 GB

Part of the INF2 Accelerated Computing family — Hardware accelerators (GPUs, FPGAs) for specialized compute.

Name:infAWS Inferentia Acceleratorsinf = AWS Inferentia AcceleratorsPurpose-built AWS Inferentia silicon designed for ultra-low latency, high-throughput machine learning inference.22nd Generation2 = 2nd GenerationAWS 2nd generation hardware platform with updated CPU/system architecture.8xlarge8 Extra Large8xlarge = 8 Extra Large32 vCPUs • 128.0 GiB RAM • EBS only • Up to 25 Gigabit Network • 1x GPU (32.0 GB VRAM)
Price (Linux)
$1.9679/ hr
~$1,436.54/mo
Compute
32vCPUs
16 cores • x86_64
Memory
128.0GiB
4.0 GiB / vCPU
Value
$0.0615/ vCPU-hr
$0.0154/GiB-hr
Network
25Gbps
Baseline: 16.667 Gbps
Storage & EBS
1,250MB/s
Max: 40,000 IOPS

Technical Specifications & Limits

Hardware specs, container thresholds, and system limits

Compute & Processor

ProcessorAMD EPYC 7R13 Processor
Architecturex86_64 (x86_64)
vCPUs / Threads32 vCPUs
Physical Cores16
Clock Speed2.95 GHz
Intel AVX / AVX2No
Intel Turbo BoostNo
Bare MetalNo (Virtualized)

Performance & Benchmarks

CoreMark Score692,341
CoreMark / vCPU21,635.7 / core
FFmpeg Transcoding198 FPS
CoreMark per Dollar351,824 score / $
Cost per vCPU-hr$0.0615
Cost per GiB-hr$0.0154

Containers & Kubernetes

EKS Max Pods234 pods
Max ECS Tasks90 tasks
Max ENIs8
IPv4/v6 per ENI30
Total Addressable IPs240 total IPs
CNI Prefix DelegationSupported (High Density)

Networking & Bandwidth

Network PerformanceUp to 25 Gigabit
Baseline Bandwidth16.667 Gbps
Burst / Peak Bandwidth25 Gbps
Enhanced NetworkingYes
EFA SupportNo
IPv6 SupportNative Dual-Stack Supported

Storage & EBS Optimization

Storage ConfigurationEBS only
Local NVMe DiskNone (EBS only)
EBS Baseline Throughput1250 MB/s
EBS Max Throughput1250 MB/s
EBS Baseline IOPS40,000
EBS Max / Burst IOPS40,000
Dedicated EBS Throughput10 Gbps

GPU & Hardware Acceleration

GPU ModelNVIDIA Accelerator
Accelerator Count1x GPU(s)
Total GPU Memory32.0 GB
Memory ArchitectureGDDR / HBM
CUDA Compute Capability—
Acceleration RuntimesCUDA, TensorRT, cuDNN, PyTorch, PyJAX

Platform & Architecture Details

Hypervisor / SystemAWS Nitro System (PCIe Offload)
Nitro EnclavesSupported (Isolated security enclaves)
Normalization Factor2.595436335
Generation StatusCurrent Generation (Recommended)
Date Introduced2023-04-13

Regional Pricing Matrix

Live on-demand rates across all global AWS cloud regions
Lowest Price Region
$1.9679/ hr
US East (N. Virginia) (us-east-1)
Highest Price Region
$3.3454/ hr
South America (Sao Paulo) (sa-east-1)
Global Fleet Average
$2.5658/ hr
Across all active regions
Regional Coverage
13regions
Available worldwide
Linux — Available in all 13 regions
Region NameRegion CodeOn-Demand Price
Asia Pacific (Mumbai)
ap-south-1$2.5582
Asia Pacific (Singapore)
ap-southeast-1$2.7550
Asia Pacific (Sydney)
ap-southeast-2$2.5582
Asia Pacific (Tokyo)
ap-northeast-1$2.9518
EU (Frankfurt)
eu-central-1$2.9518
EU (Ireland)
eu-west-1$2.4598
EU (London)
eu-west-2$2.9518
EU (Paris)
eu-west-3$2.7550
EU (Stockholm)
eu-north-1$2.1646
South America (Sao Paulo)Highest
sa-east-1$3.3454
US East (N. Virginia)Lowest
us-east-1$1.9679
US East (Ohio)Lowest
us-east-2$1.9679
US West (Oregon)Lowest
us-west-2$1.9679