Instances›AWS EC2›i8g›i8g.metal-24xl
Compute
vCPUs96
Memory768 GiB
Mem / vCPU8.00 GiB
ProcessorGraviton4 Processor
Clock speed2.7 GHz
Architecturearm64
Storage
Type6 × 3,750 GB NVMe SSD
Bandwidth30,000 Mbps
Throughput3750 MB/s
IOPS120,000
Networking
PerformanceUp to 56.25 Gigabit
Max ENIs15
IPv6Yes
GPUNone
Family sizes
i8g.large2 vCPU · 16 GiB
i8g.xlarge4 vCPU · 32 GiB
i8g.2xlarge8 vCPU · 64 GiB
i8g.4xlarge16 vCPU · 128 GiB
i8g.8xlarge32 vCPU · 256 GiB
i8g.12xlarge48 vCPU · 384 GiB
i8g.16xlarge64 vCPU · 512 GiB
i8g.24xlarge96 vCPU · 768 GiB
i8g.metal-24xl96 vCPU · 768 GiB
i8g.48xlarge192 vCPU · 1536 GiB
i8g.metal-48xl192 vCPU · 1536 GiB
i8g.metal-24xl

The i8g.metal-24xl pairs 96 vCPU with 768 GiB of RAM, the 8.00 GiB-per-vCPU ratio that marks a storage-optimized instance. In the name, "i" is I/O, meaning local storage, "8" is the generation, "g" marks AWS Graviton (ARM) processors and "metal-24xl" is the size within the i8g family. It runs on AWS Graviton4 Processor at 2.7 GHz, and AWS introduced it in December 2024. i8g is the current generation of the i family, so there is no newer AWS size to migrate to. Spot's node ladder tops out below this size, so i8g.metal-24xl becomes 6 × mh.vs1.2xlarge (96 vCPU, 768 GiB combined), $8.24/hr on AWS in US East (N. Virginia), from $0.126/hr on Rackspace Spot. Spot delivers compute as managed Kubernetes, so a node pool of that size is the native shape of the answer rather than a workaround — provided the workload scales horizontally.

Match qualityVirtualised, not bare metal

i8g.metal-24xl is a full physical server — no virtualisation layer in the way, direct access to the hardware, and a socket count that licensing can verify.

Rackspace Spot no longer sells bare metal; everything in the catalog is a virtual machine. You get the same cores and memory (6 × mh.vs1.2xlarge, 96 vCPU, 768 GiB), but not the dedicated hardware. If you picked i8g.metal-24xl because you needed the real machine — nested virtualisation, per-socket licensing, low-level profiling — this is not a replacement at any price. If you picked it because it is the biggest i8g, it is the same compute for a lot less.

Runs as a fleetNo single Spot node has 96 vCPU; the largest comparable one has 16. So this runs as 6 nodes. That works if your workload already spreads across machines, and doesn't if one process needs all the cores together.
x86_64 onlyi8g.metal-24xl is Graviton (arm64) and Rackspace Spot has no arm64 nodes, so you'd be moving to x86. You'll need x86 builds of your images, and per-core speed differs between the chips — worth testing before you decide how many nodes to run.
No local diski8g.metal-24xl comes with 6 × 3,750 GB NVMe SSD attached to the machine. Spot nodes use network storage only, so that disk doesn't come with it. If you were using it as fast scratch space, you'd move to a persistent volume or object storage, which are slower to read.
Pricing by region
1 hr
RegionAWS on-demandAWS Spot1-yr Reserved3-yr ReservedRackspace SpotYou save
US East, N. Virginia$8.237$1.081$0.126
$8.111 savings
US Central, Chicagono AWS region$0.126
US Central, Dallasno AWS region$0.786
Hong Kong$0.126
Sydney, Australia$9.874$0.987$0.126
$9.748 savings
London, UK$9.557$1.003$0.786
$8.771 savings
US West, San Jose$0.012
US Central, Dallas 2no AWS region$1.452
When to use it
Good for
  • ●Streaming ingest that spools to local disk
  • ●Elasticsearch and OpenSearch data nodes
  • ●NoSQL stores such as Cassandra, ScyllaDB and MongoDB
  • ●Workloads that fit inside 768 GiB of RAM
  • ●Jobs that want 6 × 3,750 GB of local NVMe scratch space
  • ●Network-heavy services — AWS rates this size at Up to 56.25 Gigabit
  • ●Anything that shards cleanly across 6 nodes, which is how it maps onto Rackspace Spot
Avoid if
  • ●The working set needs more than 768 GiB of RAM
  • ●You cannot rebuild x86-only binaries or vendor images for aarch64
  • ●You need dedicated hardware. Rackspace Spot sells virtualised compute only.
  • ●The work is single-threaded. Most of these 96 vCPU would sit idle.
  • ●GPU or FPGA acceleration is needed
  • ●Workloads that never touch local disk. The instance-store premium is wasted.
$6,013/mo on AWS on-demand against $92/mo on Rackspace Spot for the 6-node equivalent — market-priced, not list-priced.
  • 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