Spheron
Unified GPU rentals with no vendor lock-in
Last reviewed Aug 30, 2026
Spheron is a GPU cloud marketplace that aggregates capacity from certified third-party data centers, letting users deploy NVIDIA and AMD GPUs from multiple providers through a single account, dashboard and bill.
Available GPUs
Hourly on-demand pricing. Click column headers to sort.
Prices last updated: October 14, 2026
Spheron pricing by GPU
Configurations, price rank, and alternatives for one GPU at a time.
Pros & Cons
Advantages
- One account and dashboard spans several GPU clouds, avoiding per-provider lock-in
- Per-minute billing past a short minimum runtime, with no contract or long-term commitment
- Wide NVIDIA range from RTX 4090 through B300 plus the AMD MI300X, with GB200, GB300 and R100 available to reserve
- Spot instances run the same hardware at up to 50% off dedicated rates
- Storage, bandwidth and root access are included in the instance rate, with a dedicated IP on NVIDIA instances
Limitations
- Capacity is supplied by third-party data centers, so hardware, regions and operational experience vary by partner
- Spot instances are interruptible and can be reclaimed when demand returns
- Large clusters, custom configurations and reserved capacity go through sales rather than self-service
- The newest accelerators (GB200, GB300, R100) are reservation or pre-order only, not on-demand
- Every instance carries a minimum runtime of 20 minutes, and a 2x AMD MI300X instance carries 60 minutes and can only be destroyed, not stopped
Key Features
Multi-provider marketplace
Access GPU capacity from multiple cloud providers and certified data centers from one account, without separate signups per provider
Per-minute billing
Instances are billed by the minute rather than rounded up to the hour, past a 20-minute minimum runtime (60 minutes for a 2x AMD MI300X)
Fast provisioning
NVIDIA instances are typically ready in under two minutes with no approval workflow or provisioning queue
Unified billing and cost tracking
Compute spend across every connected provider is tracked in a single dashboard
Full root access instances
Each VM or bare metal instance includes NVMe SSD storage, network bandwidth and full root access, with a dedicated IP on NVIDIA instances
Supplier matchmaking
Spheron sources large clusters, specific hardware and InfiniBand configurations from its partner data center network, with a typical quote turnaround of 24-48 hours
Planning tools
A GPU recommender for Hugging Face models and a training cost calculator for estimating GPU cost and time
Pricing Options
| Option | Details |
|---|---|
| On-Demand Dedicated Instances | Non-interruptible instances covered by an uptime SLA and billed per minute, with no contract or long-term commitment |
| Spot Instances | Idle capacity on the same hardware at up to 50% off dedicated rates, interruptible when the provider reclaims it |
| Reserved Commitments | Commit to a duration for locked capacity, volume pricing and dedicated support |
| Custom Clusters | Quoted deployments from 8 to 512+ GPUs with specific hardware and InfiniBand configurations sourced from partner data centers |
Availability & Support
Regions
Capacity sourced from vetted Tier 3/4 partner data centers spanning multiple continents; region is selected at deploy time
Support
Documentation, API reference and changelog at docs.spheron.ai, plus sales contact and call scheduling for reserved capacity and custom cluster quotes
Getting Started
- 1
Pick your GPU
Browse available GPU models with live pricing and filter by VRAM, architecture or price
- 2
Configure and launch
Choose a region, storage and OS image, then deploy; instances are provisioned in around two minutes
- 3
Build, train, iterate
SSH in and start working, switch to another GPU model at any time, or spin down to stop per-minute billing
- 4
Request a quote for larger needs
Submit requirements for reserved capacity or custom clusters and Spheron sources, negotiates and sets up the deployment
Frequently Asked Questions
What GPU types does Spheron offer?
Spheron offers various GPU types including A100 SXM, H100 SXM, H100 SXM, GH200, H200, H200, RTX 4090, RTX 6000 Ada, HGX B300, HGX B300, L40S, RTX 5090, B200, B200, MI300X, RTX PRO 6000, RTX PRO 6000. Check the pricing table above for current availability and pricing.
How do I get started with Spheron?
Pick your GPU, Configure and launch, Build, train, iterate, Request a quote for larger needs
What are Spheron's main advantages?
Spheron's main advantages include: One account and dashboard spans several GPU clouds, avoiding per-provider lock-in, Per-minute billing past a short minimum runtime, with no contract or long-term commitment, Wide NVIDIA range from RTX 4090 through B300 plus the AMD MI300X, with GB200, GB300 and R100 available to reserve, Spot instances run the same hardware at up to 50% off dedicated rates, Storage, bandwidth and root access are included in the instance rate, with a dedicated IP on NVIDIA instances.
What are Spheron's limitations?
Spheron's main limitations include: Capacity is supplied by third-party data centers, so hardware, regions and operational experience vary by partner, Spot instances are interruptible and can be reclaimed when demand returns, Large clusters, custom configurations and reserved capacity go through sales rather than self-service, The newest accelerators (GB200, GB300, R100) are reservation or pre-order only, not on-demand, Every instance carries a minimum runtime of 20 minutes, and a 2x AMD MI300X instance carries 60 minutes and can only be destroyed, not stopped.
Compare Providers
Find the best prices for the same GPUs and models from other providers
