Spot instances
Discounted cloud compute that can be reclaimed at short notice — ideal for fault-tolerant, interruptible workloads.
Updated 2026-04-22 · 3 min read
Definition
Spot instances are spare cloud capacity sold at steep discounts in exchange for the provider's right to reclaim them on short notice. On AWS this is EC2 Spot; Azure has Spot VMs; GCP has Spot VMs and Preemptible VMs.
Why it matters
Discounts of 60–90% off on-demand turn Spot into the default compute tier for any workload that can tolerate occasional termination. Batch processing, CI/CD, rendering, ML training, stateless web scale-out — all are natural fits.
When to use
- Batch and ETL jobs that can restart from checkpoints.
- CI/CD agents and ephemeral runners.
- Stateless web services behind a load balancer with fast replacement.
- ML training jobs that checkpoint frequently.
- Avoid for: stateful single-instance services, long-running jobs without checkpointing, or anything with strict uptime SLAs that can't be met by diverse-pool configurations.
Related Terms
Savings Plans
A flexible commitment discount covering a steady compute spend rate (USD/hr) across instance families, regions, and sometimes services.
Reserved Instance
A commitment-based discount on cloud compute in exchange for a 1- or 3-year term — typically 30–60% off on-demand pricing.
Right-sizing
Matching compute capacity to actual workload demand to eliminate over-provisioning.