Spot Instances

Same Hardware.
Up to 50% Off.

Run batch jobs, training experiments, and non-urgent workloads on discounted spot instances. Same NVIDIA GPUs, same performance, lower price. Instances can be terminated when demand spikes.

Spot Savings
Up to 50%Cost Savings
SameHardware
Per-MinBilling

Spot instances use idle GPU capacity from our data center network. Same hardware, same data centers, same performance. The only difference is the price tag and the possibility of interruption during demand spikes.

How Spot Instances Work

01

Idle Capacity

GPUs not currently rented on-demand become available as spot instances at a significantly discounted rate.

02

You Deploy

Launch a spot instance just like an on-demand one. Same configuration, same GPU, lower price.

03

You Run Your Job

Your workload runs at full performance. No throttling. Checkpoint regularly for fault tolerance.

04

Termination

If on-demand demand rises, your spot instance is terminated immediately. Checkpoint regularly so you can resume from where you left off.

Built for These Workloads

Any workload that can tolerate interruption is a candidate for spot pricing. Checkpoint your work, and restarts are painless.

Hyperparameter Tuning

Run hundreds of short experiments at a fraction of the cost. Each run is independent, so interruptions only affect one trial.

Data Preprocessing

ETL pipelines and dataset transformations that can be restarted from the last checkpoint. Perfect for spot pricing.

Model Evaluation

Run inference benchmarks and validation passes on cheaper hardware. Results don't change based on instance type.

Research and Prototyping

Explore architectures and ideas without worrying about burning through your GPU budget. Iterate fast, spend less.

Spot vs On-Demand

Pricing
On-DemandStandard per-minute rates
SpotUp to 50% off on-demand
Availability
On-DemandGuaranteed while running
SpotSubject to capacity
Best for
On-DemandProduction, critical training
SpotBatch jobs, experiments
Billing
On-DemandPer-minute
SpotPer-minute
GPU Models
On-DemandAll models
SpotSelect models (availability varies)
Interruption
On-DemandNever
SpotCan be terminated immediately
Hybrid Strategy

Mix Spot and On-Demand

Run your critical training loop on an on-demand instance while offloading evaluation jobs to spot. Mix instance types across your pipeline to optimize cost without sacrificing reliability where it matters.

Critical training on guaranteed on-demand capacity
Evaluation and validation on discounted spot instances
Data preprocessing on spot for cost-efficient pipelines
Pipeline View
On-Demand
Training (Critical)
Live
Spot (50% off)~50% Savings
Evaluation
Preprocessing
Experiments

Compare Instance Types

Spot Instance FAQ

What are spot GPU instances?

Spot instances let you rent GPUs at a discount by using idle capacity in our data center network. They run on the same hardware as on-demand instances but cost up to 50% less. The tradeoff: your instance can be terminated when demand for that GPU model increases.

How much can I save with spot instances?

Savings vary by GPU model and current demand. Typical discounts range from 20% to 50% off on-demand pricing. Check the pricing page for current spot rates on each GPU model.

View pricing

What happens when a spot instance is terminated?

Spot instances can be terminated immediately when on-demand demand increases. There is no grace period, so design your pipeline with regular checkpointing. Save progress frequently, and restarts pick up from the last checkpoint.

Which workloads are a good fit for spot instances?

Any workload that handles interruption gracefully: hyperparameter searches, data preprocessing, model evaluation, batch inference, and research experiments. If you can checkpoint and restart, you can save money with spot.

Can I combine spot and on-demand instances?

Yes. Many teams run critical training jobs on on-demand instances while offloading evaluation, preprocessing, and experimental runs to spot instances. This hybrid approach gives you reliability where it counts and savings everywhere else.

GPU Infrastructure

Cut Your GPU Bill in Half

Same GPUs. Same data centers. Same performance. If your workload can tolerate interruption, there's no reason to pay full price.

Deploy Time
< 60s
Uptime SLA
99.9%
GPU Models
10+
Starting At
$0.58/hr