LIVE — 30-min refresh

GPU Price Alerts
Never Miss a Spot Drop

Set a price threshold for any GPU on any cloud. Get an email the moment it dips below.

Current H100 spot: $2.06/hr on CoreWeave
How it works
Three steps. No credit card required.
01

Pick your GPU

Choose from H100, A100, L40S, A10G, T4, RTX 4090, and more. Filter by provider if you have a preference.

02

Set your threshold

Enter the price per hour you're willing to pay. We'll watch every 30 minutes across all 7 providers.

03

Get the email

When spot prices drop below your threshold, you get a notification immediately — while the deal lasts.

GPU spot prices swing hard
7-day price range across GPU models — alerts catch these moves in real time.
Loading...
Price data populating
H100 spot prices right now
Set an alert below your target — we'll tell you when any provider hits it.
Provider Current Spot Action
★ CoreWeave $2.06/hr Alert me
Azure $2.13/hr Alert me
Lambda Cloud $2.49/hr Alert me
RunPod $2.59/hr Alert me
Vast.ai $4.40/hr Alert me
Google Cloud $26.77/hr Alert me
AWS $60.12/hr Alert me
Cheapest available spot per provider
Live prices — updated every 30 minutes from 7 cloud APIs.
AWS
NVIDIA T4
$0.09/hr
Set alert →
Azure
NVIDIA T4
$0.10/hr
Set alert →
CoreWeave
NVIDIA A100 SXM
$1.19/hr
Set alert →
Google Cloud
NVIDIA T4
$0.16/hr
Set alert →
Lambda Cloud
NVIDIA A10G
$0.59/hr
Set alert →
RunPod
NVIDIA V100
$0.19/hr
Set alert →
Vast.ai
NVIDIA RTX 3090
$0.01/hr
Set alert →
Common questions
Everything about GPU price alerts on RoofRun.
Which GPUs can I set price alerts for?
H100, A100 (40GB and 80GB), L40S, A10G, T4, and RTX 4090. We track 16+ GPU models across 7 providers. If a GPU has live spot pricing in our system, you can alert on it.
Which cloud providers are supported?
AWS, Google Cloud, Azure, Lambda Cloud, CoreWeave, RunPod, and Vast.ai. You can set a provider-specific alert or watch across all providers at once. See individual provider pages for current availability.
How often do you check prices?
Every 30 minutes, pulling directly from each provider's live API. GPU spot prices can move fast — a 30-minute poll window is tight enough to catch most meaningful drops while staying within API rate limits.
How many alerts can I set?
Unlimited alerts on any GPU, any provider, any threshold. Alerts trigger once per 6 hours per GPU/provider pair — so you won't get spammed during volatile periods.
Is it free?
Yes. Create a free account at /app, no credit card needed. Premium plans add historical data exports and API access if you need them.
Why do H100 spot prices vary so much between providers?
Each provider sets spot prices based on their own supply/demand. RunPod and Lambda Cloud typically run cheaper than AWS and CoreWeave because they have more spare capacity. The Trends page shows 30-day stability ratings per provider.
Explore more
Tools for finding the cheapest GPU cloud compute.

Start tracking GPU spot prices

Free account. Set unlimited alerts. Get notified when prices drop.

Create Free Account →