I checked the CPU and CloudTrail information of EC2 C9g instances that became available in the Tokyo region

Introduction

On September 3, 2026, EC2 C9g and C9gd instances powered by AWS Graviton5 became available in 4 regions, including the Tokyo region (ap-northeast-1).

https://aws.amazon.com/jp/about-aws/whats-new/2026/09/ec2-c9g-c9gd-asia-pacific-tokyo/

On the same day, the general-purpose M9g and M9gd also became available in the Tokyo region.

https://dev.classmethod.jp/articles/ec2-m9g-graviton5-tokyo-launch/

In this article, I launched a c9g.medium in the Tokyo region, confirmed the launch in CloudTrail, and retrieved CPU information.

Verification Details

Launching the Instance

I launched a c9g.medium using the AWS CLI.

aws ec2 run-instances \
  --region ap-northeast-1 \
  --image-id ami-0abcde8261ecfaf82 \
  --instance-type c9g.medium \
  --subnet-id subnet-xxxxx \
  --iam-instance-profile Name=ec2-ssm-core \
  --associate-public-ip-address \
  --metadata-options "HttpTokens=required,HttpEndpoint=enabled" \
  --tag-specifications 'ResourceType=instance,Tags=[{Key=Name,Value=c9g-tokyo-verify}]' \
  --query 'Instances[0].InstanceId' \
  --output text

i-0123456789abcdef0

Confirming the Launch in CloudTrail

After launching, I confirmed the RunInstances event in CloudTrail.

aws cloudtrail lookup-events \
  --lookup-attributes AttributeKey=EventName,AttributeValue=RunInstances \
  --start-time "2026-09-05T06:29:00+09:00" \
  --end-time "2026-09-05T06:30:00+09:00" \
  --region ap-northeast-1 \
  --query 'Events[0].CloudTrailEvent' \
  --output text | jq '.'

Below is information excerpted from the launch event.

{
  "eventTime": "2026-09-04T21:29:16Z",
  "eventName": "RunInstances",
  "awsRegion": "ap-northeast-1",
  "requestParameters": {
    "instanceType": "c9g.medium",
    "instancesSet": {
      "items": [{ "imageId": "ami-0abcde8261ecfaf82" }]
    }
  },
  "responseElements": {
    "instancesSet": {
      "items": [
        {
          "instanceId": "i-0123456789abcdef0",
          "instanceType": "c9g.medium",
          "placement": {
            "availabilityZone": "ap-northeast-1a"
          },
          "architecture": "arm64",
          "launchTime": 1788557356000
        }
      ]
    }
  }
}

I was able to confirm that the instanceType: c9g.medium specified in the CLI was launched in ap-northeast-1a and recorded in CloudTrail.

Verification Environment

I launched one c9g.medium in ap-northeast-1a. The OS is Amazon Linux 2023.

instance-type: c9g.medium
availability-zone: ap-northeast-1a
ami-id: ami-0abcde8261ecfaf82
kernel: 6.18.44-99.149.amzn2023.aarch64
PRETTY_NAME="Amazon Linux 2023.12.20260831"

CPU Information

I ran lscpu. Below is an excerpt of the key points from the output.

Architecture:                            aarch64
BIOS Vendor ID:                          AWS
BIOS Model name:                         AWS Graviton5
Stepping:                                r0p1
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti ecv afp wfxt
L1d cache:                               64 KiB (1 instance)
L1i cache:                               64 KiB (1 instance)
L2 cache:                                2 MiB (1 instance)
L3 cache:                                48 MiB (1 instance)

All listed items matched the values obtained in the M9g article. Since the C9g and M9g are equipped with the same Graviton5 processor, there is no difference in CPU information.

Architecture:                            aarch64
CPU op-mode(s):                          64-bit
Byte Order:                              Little Endian
CPU(s):                                  1
On-line CPU(s) list:                     0
Vendor ID:                               ARM
BIOS Vendor ID:                          AWS
BIOS Model name:                         AWS Graviton5
Model:                                   1
Thread(s) per core:                      1
Core(s) per socket:                      1
Socket(s):                               1
Stepping:                                r0p1
BogoMIPS:                                2000.00
Flags:                                   fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti ecv afp wfxt
L1d cache:                               64 KiB (1 instance)
L1i cache:                               64 KiB (1 instance)
L2 cache:                                2 MiB (1 instance)
L3 cache:                                48 MiB (1 instance)
NUMA node(s):                            1
NUMA node0 CPU(s):                       0

I retrieved processor information using dmidecode.

Processor Information
	Family: ARMv9
	Manufacturer: AWS
	ID: 41 D8 0F 41 00 00 00 00
	Signature: Implementor 0x41, Variant 0x0, Architecture 15, Part 0xd84, Revision 1
	Version: AWS Graviton5
	Max Speed: 3300 MHz
	Current Speed: 3300 MHz
	Core Count: 1

The CPU part is 0xd84, which was the same value as the M9g.

# dmidecode 3.6
Getting SMBIOS data from sysfs.
SMBIOS 3.0.0 present.

Handle 0x0004, DMI type 4, 42 bytes
Processor Information
	Socket Designation: CPU0
	Type: Central Processor
	Family: ARMv9
	Manufacturer: AWS
	ID: 41 D8 0F 41 00 00 00 00
	Signature: Implementor 0x41, Variant 0x0, Architecture 15, Part 0xd84, Revision 1
	Version: AWS Graviton5
	Voltage: Unknown
	External Clock: Unknown
	Max Speed: 3300 MHz
	Current Speed: 3300 MHz
	Status: Populated, Enabled
	Upgrade: None
	L1 Cache Handle: Not Provided
	L2 Cache Handle: Not Provided
	L3 Cache Handle: Not Provided
	Serial Number: AWS Graviton5
	Asset Tag: AWS Graviton5
	Part Number: AWS Graviton5
	Core Count: 1
	Core Enabled: 1
	Thread Count: 1
	Characteristics: None

I confirmed the SVE vector length.

cat /proc/sys/abi/sve_default_vector_length

16

The vector length is 16 bytes (128 bits), which is no different from the M9g.

processor	: 0
BogoMIPS	: 2000.00
Features	: fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics fphp asimdhp cpuid asimdrdm jscvt fcma lrcpc dcpop sha3 asimddp sha512 sve asimdfhm dit uscat ilrcpc flagm sb paca pacg dcpodp sve2 sveaes svepmull svebitperm svesha3 flagm2 frint svei8mm svebf16 i8mm bf16 dgh rng bti ecv afp wfxt
CPU implementer	: 0x41
CPU architecture: 8
CPU variant	: 0x0
CPU part	: 0xd84
CPU revision	: 1

Machine (1797MB total)
  Package L#0
    NUMANode L#0 (P#0 1797MB)
    L3 L#0 (48MB) + L2 L#0 (2048KB) + L1d L#0 (64KB) + L1i L#0 (64KB) + Core L#0 + PU L#0 (P#0)
  HostBridge
    PCI 00:04.0 (NVMExp)
      Block(Disk) "nvme0n1"
  HostBridge
    PCIBridge
      PCIBridge
        PCIBridge
          PCI 27:00.0 (Ethernet)
            Net "ens34"

Pricing

I retrieved the Tokyo region pricing via the Pricing API.

Comparing the ap-northeast-1 Linux On-Demand pricing as of September 5, 2026 against C8g, the price was +9.0% across all sizes listed in the table. Monthly cost is calculated based on 720 hours.

Comparing C9g pricing between the Tokyo region and us-east-1, Tokyo pricing was +25.4% across all sizes listed in the table.

Comparing the compute-optimized C9g and the general-purpose M9g, both equipped with the same Graviton5, C9g pricing is approximately 14% cheaper than M9g across all sizes listed in the table.

The C9g comes with 2 GB of memory per vCPU, which is half the 4 GB/vCPU of the M9g. For workloads that prioritize CPU performance over memory, choosing the C9g can help reduce costs.

Summary

I launched a compute-optimized c9g.medium in ap-northeast-1a in the Tokyo region. I confirmed the launch in CloudTrail and verified that the CPU information matches that of the general-purpose M9g.

The Linux On-Demand pricing for C9g is 9% higher than C8g, but AWS states that C9g delivers up to 25% improved compute performance compared to C8g powered by Graviton4. Also, compared to the M9g with the same Graviton5, C9g is approximately 14% cheaper, making it a cost-efficient choice for workloads that prioritize CPU performance over memory.

With this, both the general-purpose M9g and the compute-optimized C9g are now available in the Tokyo region. Once the memory-optimized R9g is added to the lineup, Graviton5 will be available for major workload types. I look forward to R9g becoming available in the Tokyo region.