The m5d.24xlarge pairs 96 vCPU with 384 GiB of RAM, the 4.00 GiB-per-vCPU ratio that marks a general-purpose instance. In the name, "m" is the general-purpose line, "5" is the generation, "d" marks local NVMe storage and "24xlarge" is the size within the m5d family. It runs on Intel Xeon Platinum 8175 at 3.1 GHz, and AWS introduced it in June 2018. AWS has since shipped m6i, m7i, m8i and m9g; m6i.24xlarge is the direct replacement at this size, scoring 1,587,761 CoreMark/s against this instance's 1,386,432 — roughly 15% more throughput for the same vCPU count. Spot's node ladder tops out below this size, so m5d.24xlarge becomes 6 × gp.vs1.2xlarge (96 vCPU, 384 GiB combined), about 95% below the US East (N. Virginia) on-demand rate. 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 qualityScale-out match
m5d.24xlarge gives you 96 vCPU and 384 GiB in a single machine. The largest comparable Spot node, gp.vs1.2xlarge, is 16 vCPU and 64 GiB, so the equivalent here is 6 nodes rather than one.
Spot runs as managed Kubernetes, so a 6-node pool is the normal way to run things rather than a workaround. If your workload already sits behind a load balancer, a queue or a scheduler, it spreads across them without changes. If it needs all 96 cores and 384 GiB in one process, it does not.
No local diskm5d.24xlarge comes with 4 × 900 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.
