AWS EC2 Instances›Compute Optimized›c8a›c8a.48xlarge

Compared with Rackspace Spot

AWSRackspace Spot
Machinec8a.48xlarge12 × ch.vs1.2xlarge
vCPU192192
Memory384 GiB384 GiB
Local diskNoneNone
On-demand / hr$10.346$3.132
Spot / hr$2.777$0.492
Prices in US East, N. Virginia

Compute

vCPUs192
Memory384 GiB
Mem / vCPU2.00 GiB
ProcessorEPYC 9R45 Processor
Clock speedUp to 4.5 GHz
Architecturex86_64
GPU0
GPU Wattage0 W
GPU ArchitectureNone
Video Memory0 GiB
FPGA0
ffmpeg FPS410
CoreMark / sec6,834,509

Storage

TypeEBS only
Bandwidth60,000 Mbps
Throughput7500 MB/s
IOPS240,000
Base Bandwidth60,000 Mbps
Base Throughput7500 MB/s
Base IOPS240,000
Devices0
NVMe DriveNo
SSDNo
Swap PartitionNo

Networking

Performance75 Gigabit
Max ENIs24
IPv6Yes
GPUNone
Placement GroupSupported
TrunkingYes
Branch Interface98
IPs / ENI64
Baseline Bandwidth75 Gbps

NUMA Architecture

Uses NUMAYes
Node Count2
Max Distance32
Cores / Node96
Threads / Node96
Memory / Node189,892 MB
L3 Cache / Node384 MB
L3 Cache SharedYes

AWS Metadata

Instance typec8a.48xlarge
Full nameC8A 48xlarge
FamilyCompute optimized
Variantc8a
GenerationCurrent
Bare MetalNo
EMR supportNo
DedicatedSupported
Introduced2025-12-02
Family sizes

c8a.48xlarge

AWS's c8a.48xlarge is a compute-optimized instance with 192 vCPU and 384 GiB. In US East (N. Virginia), the closest Rackspace Spot server is ch.vs1.2xlarge at 16 vCPU and 32 GiB, so you will need 12 of them to get a total of 192 vCPU and 384 GiB.

Even with 12 servers, Rackspace Spot's on-demand price is about 70% below AWS on-demand and its Spot market price is about 82% below AWS Spot. Rackspace Spot offers managed Kubernetes, so you can run them together in one cluster.

Server comparison caveats
Scale-out match

The AWS c8a.48xlarge gives you 192 vCPU and 384 GiB in a single machine. The largest comparable Rackspace Spot server, ch.vs1.2xlarge, is 16 vCPU and 32 GiB, so the equivalent here is 12 servers rather than one.

This setup works best for Kubernetes workloads, or for workloads that already sit behind a load balancer, a queue or a scheduler, because they spread across the 12 servers without changes. If it needs all 192 cores and 384 GiB in one process, it does not.

Pricing by region

1 hr

Regions are Rackspace Spot locations.

RegionAWS on-demandAWS Spot1-yr Reserved3-yr ReservedRackspace on-demandRackspace SpotOn-demand price savingsSpot price savings
US East, N. Virginia$10.346$2.777not offerednot offered$3.132$0.492
$7.214 (70%)
$2.285 (82%)
US Central, Chicagono AWS regionn/an/an/a$3.324$0.492n/an/a
US Central, Dallasno AWS regionn/an/an/a$3.300$1.572n/an/a
Hong Kongnot sold heren/an/an/a$3.240$0.132n/an/a
Sydney, Australianot sold heren/an/an/a$3.504$0.432n/an/a
London, UKnot sold heren/an/an/a$3.108$0.492n/an/a
US West, San Josenot sold heren/an/an/a$3.120$0.024n/an/a
US Central, Dallas 2no AWS regionn/an/an/a$3.120$1.464n/an/a
US East, N. Virginia 2$10.346$2.777not offerednot offered$3.120$1.464
$7.226 (70%)
$1.313 (47%)

Prices updated 18 Nov 2026.

When to use it

Good for
  • ●Video encoding and transcoding
  • ●High-traffic web and API serving
  • ●Scientific modelling and simulation
  • ●Workloads that fit inside 384 GiB of RAM
  • ●Parallel work that can occupy 192 vCPU
  • ●Network-heavy services (AWS rates this size at 75 Gigabit)
  • ●Anything that shards cleanly across 12 nodes, which is how it maps onto Rackspace Spot
Avoid if
  • ●The working set needs more than 384 GiB of RAM
  • ●Memory-hungry applications. A memory-optimized family gives more RAM per dollar.
  • ●More than 2.00 GiB of RAM per vCPU is needed, since that ratio is fixed for the whole family
  • ●The work is single-threaded. Most of these 192 vCPU would sit idle.
  • ●Local NVMe or instance-store disk is a hard requirement
  • ●GPU or FPGA acceleration is needed
c8a.48xlarge is $7,553/mo on AWS on-demand and $2,286/mo on Rackspace Spot on-demand for 12 servers, 70% less. At the Spot market price, the same capacity is $359/mo.
  • Rackspace Spot sells spare capacity at market prices. It is the cheapest compute on the internet.
  • Prices are set by supply and demand. You pay what the market decides.
  • Compute is delivered as fully managed Kubernetes. All infrastructure is taken care of.
  • Egress costs $0.01/GB after the free allowance. AWS charges $0.09/GB.
Build your Cloudspace

Frequently asked questions

How much does c8a.48xlarge cost?

The AWS c8a.48xlarge costs $10.35/hr on-demand in US East (N. Virginia), about $7,553/mo running 24/7, or $2.78/hr at the AWS Spot price. On Rackspace Spot, 12 × ch.vs1.2xlarge is $3.13/hr on-demand, or $0.492/hr at the Spot price. Rackspace Spot also sells in Chicago and Dallas, where AWS operates no region at all, so those rows have no AWS price to compare against.

How much is c8a.48xlarge per month?

At 730 hours a month, c8a.48xlarge in US East (N. Virginia) is $7,553/mo on AWS on-demand, or $2,027/mo at the AWS Spot price. The same capacity on Rackspace Spot (12 × ch.vs1.2xlarge) is $2,286/mo on-demand, or $359/mo at the Spot price. Rackspace Spot prices move with demand, so the Spot figure is a typical rate rather than a fixed one.

How many Rackspace Spot nodes replace one c8a.48xlarge?

12. The ch.vs1.2xlarge server is 16 vCPU and 32 GiB, and Rackspace Spot's whole VM catalog tops out at 16 vCPU, so covering 192 vCPU and 384 GiB takes 12 of them: 192 vCPU and 384 GiB in total. Per-region fleet sizes differ and are shown in the table.

What is the difference between c8a and c9g?

c9g is the newer generation. It runs on AWS Graviton5 Processor against c8a's AMD EPYC 9R45 Processor. On-demand in US East (N. Virginia) the newer size is 19% cheaper. Neither generation changes the Rackspace Spot side: both map to the same Rackspace Spot servers.

What is a cheaper alternative to c8a.48xlarge?

Rackspace Spot 12 × ch.vs1.2xlarge (192 vCPU, 384 GiB) is $2,286/mo on-demand or $359/mo at the Spot price, against $7,553/mo on-demand or $2,027/mo at the Spot price for c8a.48xlarge on AWS in US East (N. Virginia). The trade is real: Rackspace Spot capacity is priced by demand and can be reclaimed when demand rises, and workloads have to be containerized.