Service comparison / compute pricing
Cloud Compute Pricing 2026: AWS EC2 vs Azure VMs vs GCP Compute Engine
This is the ledger: on-demand list pricing for general purpose, compute-optimised, memory-optimised, ARM, and GPU instances across all three providers in their major US region. Checked 9 November 2026 against the AWS EC2 price list, the Azure Retail Prices API and the GCP machine-type pricing pages.
Headline finding
On-demand list pricing for equivalent compute specs differs by less than 5% across providers in most US regions. The real lever is discount strategy, not list shopping.
General purpose, on-demand
| Spec | AWS | Azure | GCP |
|---|---|---|---|
| 4 vCPU / 16 GB | $0.2016/hr m7i.xlarge | $0.1920/hr D4s v5 | $0.1942/hr n2-standard-4 |
| 8 vCPU / 32 GB | $0.4032/hr m7i.2xlarge | $0.3840/hr D8s v5 | $0.3884/hr n2-standard-8 |
| 16 vCPU / 64 GB | $0.8064/hr m7i.4xlarge | $0.7680/hr D16s v5 | $0.7768/hr n2-standard-16 |
| 32 vCPU / 128 GB | $1.6128/hr m7i.8xlarge | $1.5360/hr D32s v5 | $1.5536/hr n2-standard-32 |
| 64 vCPU / 256 GB | $3.2256/hr m7i.16xlarge | $3.0720/hr D64s v5 | $3.1072/hr n2-standard-64 |
Compute optimised, on-demand
| Spec | AWS | Azure | GCP |
|---|---|---|---|
| 4 vCPU / 8 GB | $0.1785/hr c7i.xlarge | $0.1690/hr F4s v2 | $0.1701/hr c4-highcpu-4 |
| 8 vCPU / 16 GB | $0.3570/hr c7i.2xlarge | $0.3380/hr F8s v2 | $0.3402/hr c4-highcpu-8 |
| 16 vCPU / 32 GB | $0.7140/hr c7i.4xlarge | $0.6770/hr F16s v2 | $0.6804/hr c4-highcpu-16 |
Memory optimised, on-demand
| Spec | AWS | Azure | GCP |
|---|---|---|---|
| 4 vCPU / 32 GB | $0.2646/hr r7i.xlarge | $0.2520/hr E4s v5 | $0.2620/hr n2-highmem-4 |
| 8 vCPU / 64 GB | $0.5292/hr r7i.2xlarge | $0.5040/hr E8s v5 | $0.5241/hr n2-highmem-8 |
| 16 vCPU / 128 GB | $1.0584/hr r7i.4xlarge | $1.0080/hr E16s v5 | $1.0482/hr n2-highmem-16 |
GPU instances, on-demand
GPU pricing changes faster than other categories. Always verify directly with the provider. Linux on-demand list rates, checked 9 November 2026. Azure lists no L4 VM size, and GCP lists no single-GPU A3 size.
| Spec | AWS | Azure | GCP |
|---|---|---|---|
| 1x H100 | $6.88/hr p5.4xlarge | $6.98/hr NC40ads H100 v5 | no 1-GPU A3 size listed |
| 8x H100 | $55.04/hr p5.48xlarge | $98.32/hr ND96isr H100 v5 | $88.49/hr a3-highgpu-8g |
| A100 | $21.96/hr p4d.24xlarge (8x A100 40GB) | $3.673/hr NC24ads A100 v4 (1x A100 80GB) | $3.673/hr a2-highgpu-1g (1x A100 40GB) |
| 1x L4 | $0.8048/hr g6.xlarge | no L4 size listed | $0.7068/hr g2-standard-4 |
Custom silicon / ARM list rates, 4 vCPU / 16 GB
ARM and custom silicon
AWS Graviton 4 (m8g)
$0.1795/hr m8g.xlarge
About 11% below m7i.xlarge ($0.2016/hr) on list rate for the same 4 vCPU / 16 GB. Check your binaries and agents have ARM builds.
Azure Cobalt 100 (Dpsv6)
$0.1400/hr D4ps v6
About 27% below D4s v5 ($0.192/hr) on list rate for the same 4 vCPU / 16 GB.
GCP Tau T2A (Ampere Altra)
$0.1540/hr t2a-standard-4
About 21% below n2-standard-4 ($0.1942/hr) on list rate for the same 4 vCPU / 16 GB.
Spot / Preemptible / Spot VMs
Interruptible compute
| Provider | Discount | Interruption | Use cases |
|---|---|---|---|
| AWS Spot | up to 90% off on-demand | 2-minute warning; interruption frequency varies by instance type and region (see the AWS Spot Instance Advisor) | ECS, EKS data plane, batch, training, CI runners, Fargate Spot. |
| Azure Spot VMs | up to 90% off pay-as-you-go | 30-second warning, eviction-based on capacity or price | AKS spot pools, Azure Batch, low-priority VM scale sets. |
| GCP Spot VMs | up to 91% off on-demand | 30-second warning, no 24-hour limit | GKE spot node pools, Dataproc, Cloud Build workers. |
Region pricing variation
Same instance, different region
Region selection can be one of the largest single compute cost levers. Linux on-demand rate for a 4 vCPU / 16 GB general-purpose VM (AWS m7i.xlarge, Azure D4s v5), with the premium over each provider's US East rate. Checked 9 November 2026.
| Region | AWS | Azure | GCP |
|---|---|---|---|
| US East (N. Virginia / East US) | $0.2016/hr (baseline) | $0.192/hr (baseline) | $0.1942/hr us-central1 |
| Europe (Ireland / West Europe) | $0.2247/hr (+11%) | $0.23/hr (+20%) | not compared here |
| Singapore (ap-southeast-1 / Southeast Asia) | $0.252/hr (+25%) | $0.24/hr (+25%) | not compared here |
| Sao Paulo (sa-east-1 / Brazil South) | $0.3213/hr (+59%) | $0.306/hr (+59%) | not compared here |
Common questions
FAQ
Which cloud has the cheapest compute?+
On-demand list pricing for equivalent specs is within about 5% across AWS, Azure, and GCP in most US regions. Azure D-series and GCP n2-standard tend to list a few cents per hour below AWS m7i, but the difference is too small to drive provider selection. Effective rate after Sustained Use Discounts (GCP) or Savings Plans (AWS, Azure) varies more than list rate.
Are ARM instances cheaper?+
On list rate, yes. In the major US regions a 4 vCPU / 16 GB ARM instance lists about 11% below the x86 equivalent on AWS (m8g.xlarge $0.1795/hr against m7i.xlarge $0.2016/hr), about 27% below on Azure (D4ps v6 $0.14/hr against D4s v5 $0.192/hr) and about 21% below on GCP (t2a-standard-4 $0.154/hr against n2-standard-4 $0.1942/hr). Performance per vCPU differs by workload, so benchmark your own service. The catch is compatibility: Linux containers and most managed services run on ARM, but third-party agents, proprietary binaries, and Windows workloads do not always have ARM builds.
How much do GPU instances actually cost?+
On-demand list rates checked 9 November 2026 in the major US regions: an 8x H100 node is $55.04/hr on AWS (p5.48xlarge), $98.32/hr on Azure (ND96isr H100 v5) and $88.49/hr on GCP (a3-highgpu-8g). A single H100 is $6.88/hr on AWS (p5.4xlarge) and $6.98/hr on Azure (NC40ads H100 v5). The machines differ in CPU, memory, local storage and networking, so compare the full spec, not just the GPU. Spot and committed capacity are priced separately and change often.
Why does the same instance cost more in Sao Paulo?+
Each provider prices each region separately. Checked 9 November 2026, a 4 vCPU / 16 GB general-purpose VM lists 59% above its US East rate in Sao Paulo on both AWS (m7i.xlarge $0.3213/hr against $0.2016/hr) and Azure (D4s v5 in Brazil South $0.306/hr against $0.192/hr). Singapore is about 25% above on both, and Europe 11% (AWS Ireland) to 20% (Azure West Europe) above. For workloads where region is not a regulatory requirement, the region can be one of the largest single compute cost levers.
Should I commit to RIs or Savings Plans on compute?+
For stable production compute, both AWS and Azure offer up to 72% discount on 3-year all-upfront commitments. Compute Savings Plans on AWS and Savings Plans on Azure are flexible across families. Reserved Instances lock you to a configuration for slightly more discount. The rule of thumb: cover 60-80% of steady-state with commitments, leave the spiky top on on-demand or Spot.
Continue reading