p4d.24xlarge
$23.7157/hr
~$17312.45/month on-demand in eu-west-1
The p4d.24xlarge is an Accelerated computing instance with 96 vCPUs and 1152 GiB of memory, 8x NVIDIA A100 with up to 100 Gbps bandwidth, starting at $23.716 per hour.
Last updated: Jan 15, 2026, 12:23 PM UTC
Pricing varies by operating system
All pricing cards update to show this payment type
Pricing Options
Compute SP
Most flexible - any instance
1 Year - No Upfront
$18.0983/hr
~$13211.73/mo
Save 24%
3 Year - No Upfront
$16.2365/hr
~$11852.64/mo
Save 32%
EC2 Instance SP
Locked to family - cheaper
1 Year - No Upfront
$15.0819/hr
~$11009.77/mo
Save 36%
3 Year - No Upfront
$10.1478/hr
~$7407.90/mo
Save 57%
Reserved
Region & instance locked
1 Year - No Upfront
$15.0819/hr
~$11009.77/mo
Save 36%
3 Year - No Upfront
$10.1478/hr
~$7407.90/mo
Save 57%
Spot Pricing
Up to 90% savings - flexible workloads
Cheapest AZ
eu-west-1a
$10.9900/hr
~$8022.70/mo
Save 61% vs On-Demand
10-15%interruption rate
Price Range (24h)
Min
$9.7472
Avg
$10.2782
Max
$10.9900
All Availability Zones
eu-west-1a$10.9900/hr
eu-west-1a$10.0782/hr
eu-west-1a$9.9532/hr
Instance Specifications
Detailed hardware and capabilities for p4d.24xlarge
Processor
ManufacturerIntel
Clock Speed3 GHz
Sustained Speed3 GHz
Architecturex86_64
vCPUs96
Memory1152 GiB
Memory per vCPU12.00 GiB
Network
Performance4x 100 Gigabit
Bandwidth100 Gbps
Max Interfaces60
IPv4 per Interface50
Enhanced Networking
EFA Support
IPv6
Storage
Instance Storage8000 GB
Storage Typessd
NVMe Drive
EBS Optimized
EBS Encryption
EBS NVMe Support
Max Bandwidth19000 Mbps (2375.00 MB/s)
Max IOPS80,000
Baseline Bandwidth19000 Mbps (2375.00 MB/s)
Baseline IOPS80,000
GPU
Count8
ManufacturerNVIDIA
ModelA100
Memory320 GiB
Capabilities
FamilyP4D
GenerationCurrent
Burstable
Bare Metal
Dedicated Hosts
Hibernation
Auto Recovery
Free Tier
Usage Classescapacity-block, on-demand, spot
Root Devicesebs
Virtualizationhvm
Regional Pricing Comparison
Compare p4d.24xlarge (Linux) pricing across all regions - sorted by cost
Current region (eu-west-1)Other regions
💡 Click any bar to open that region in a new tab