Question 1

What GPUs are available for rent on Spheron?

Accepted Answer

Spheron offers a full range of NVIDIA GPUs for rent including the B300, H100, B200, H200, GH200, A100, L40S, RTX 5090, RTX PRO 6000, and RTX 4090. All GPUs are enterprise-grade with dedicated VRAM and available for instant deployment with no long-term contracts.

Question 2

How much does it cost to rent a GPU on Spheron?

Accepted Answer

GPU rental prices on Spheron start from $0.58/hr for an RTX 4090 and go up to $3.50/hr for the B300. All pricing is per-minute with no minimum commitment, no hidden fees, and no warm-up charges. Spheron is typically 57-68% cheaper than hyperscaler on-demand rates.

Question 3

How quickly can I deploy a GPU instance?

Accepted Answer

GPU instances on Spheron deploy in 60-90 seconds. You get full SSH root access with a dedicated IP address immediately after provisioning. No waiting for approval or manual setup required.

Question 4

Do I need a long-term contract to rent GPUs?

Accepted Answer

No. Spheron uses per-minute billing with no minimum commitment and no long-term contracts. You pay only for the GPU time you actually use, and you can stop at any time without penalty.

Question 5

What is the difference between H100 and A100 GPUs?

Accepted Answer

The NVIDIA H100 (Hopper architecture) offers 80 GB HBM3 memory with 3.35 TB/s bandwidth, while the A100 (Ampere architecture) has 80 GB HBM2e with 2 TB/s bandwidth. The H100 delivers roughly 3x the AI training performance of the A100 and is better suited for large-scale LLM training, while the A100 remains cost-effective for inference and smaller workloads.

Question 6

Which GPU should I choose for AI training vs inference?

Accepted Answer

For large-scale AI training, the H100 or B200 are recommended due to their high memory bandwidth and Tensor Core performance. For inference workloads, the A100, L40S, or RTX 4090 offer excellent price-to-performance ratios. The RTX 5090 is a strong choice for budget-conscious inference at scale.

Question 7

Does Spheron provide bare-metal GPU access?

Accepted Answer

Yes. Spheron provides full VM and bare-metal GPU access with root SSH access and dedicated IP addresses. Unlike container-based platforms, you get 100% of the hardware performance with no hypervisor overhead and no noisy neighbors.

GPU Rental

Browse our GPU catalog

B300

H100

B200

H200

GH200

A100

RTX PRO 6000

RTX 5090

L40S

RTX 4090

Ready to Deploy?