# Single GPU server rental: a whole GPU in a dedicated VM > Single GPU server rental: one whole GPU in a dedicated cloud VM, from RTX 5090 to B300. Root access, 20 TB bandwidth, monthly billing, delivered in under 10 minutes. Source: https://overgpu.com/products/single-gpu-servers Updated: 2026-10-23 Publisher: OverGPU (https://overgpu.com/) ## Key points - **The whole GPU, passed through:** All of the GPU's memory and compute belong to your VM. No vGPU slicing, no MIG partitions, no time-sharing. - **Dedicated vCPU, RAM and NVMe:** Each plan has its own vCPU, memory and local NVMe, from 16 vCPU and 64 GB of RAM up to 28 vCPU and 384 GB. - **Seven GPUs to choose from:** From the GeForce RTX 5090 with 32 GB at $449 a month to the B300 with 288 GB of HBM3e at $5,190. - **20 TB of bandwidth included:** 20 TB of outbound traffic per month under fair use, with no egress fees. Inbound traffic is not counted. - **Root access over SSH:** Ubuntu 22.04 or 24.04 with NVIDIA drivers, CUDA, Docker and the NVIDIA Container Toolkit preinstalled. - **Monthly, cancel any month:** A flat price paid monthly in advance from your balance. No setup fee and no cancellation fee. ## Recommended configurations - [NVIDIA H200](https://overgpu.com/gpus/nvidia-h200): $3,090 per month — 141 GB of HBM3e: 70B models in FP8 with room for the KV cache - [NVIDIA H100 SXM](https://overgpu.com/gpus/nvidia-h100): $2,190 per month — 80 GB of HBM3 for LoRA fine-tunes and mid-size models - [NVIDIA RTX PRO 6000 Blackwell](https://overgpu.com/gpus/nvidia-rtx-pro-6000): $1,390 per month — 96 GB of GDDR7 for inference, generative media and rendering ## Overview Single GPU server rental at OverGPU means one whole physical GPU, passed through to a virtual machine that only you use. The vCPU, RAM and local NVMe that come with it are dedicated to that VM, and you log in as root over SSH. You pay one flat price per month, from $449 for a GeForce RTX 5090 to $5,190 for an NVIDIA B300 with 288 GB of HBM3e. ## How single GPU server rental works Each plan is a dedicated GPU VM. The hypervisor passes the entire GPU through to your virtual machine, so the guest operating system loads the NVIDIA driver and drives the card directly. Run `nvidia-smi` inside the VM and you see the full memory of the GPU. No other tenant has a slice of it: there is no vGPU, no MIG partition and no time-sharing scheduler between your code and the hardware. Around the GPU, every single-GPU server comes with: - A public IPv4 address and an IPv6 /64, with DDoS mitigation on the public IPs. - Root access by SSH key or password, Ubuntu images with the GPU stack preinstalled, and optional software installed for you at no extra cost, described on our [OS images](https://overgpu.com/products/os-images) page. - A 99.9% monthly availability SLA for power, network and hardware, with service credits if we fall short. - Local NVMe storage that is securely erased between customers. ## The seven single-GPU plans compared Host resources grow with the plan: the higher up the table, the more vCPU, RAM, NVMe and network come with the GPU. | GPU | GPU memory | vCPU / RAM | NVMe | Network | Price/mo | | --- | --- | --- | --- | --- | --- | | NVIDIA B300 | 288 GB HBM3e | 28 / 384 GB | 3.5 TB | 25 Gbps | $5,190 | | NVIDIA B200 | 180 GB HBM3e | 26 / 256 GB | 2.75 TB | 25 Gbps | $4,590 | | NVIDIA H200 | 141 GB HBM3e | 24 / 240 GB | 2 TB | 25 Gbps | $3,090 | | NVIDIA H100 SXM | 80 GB HBM3 | 24 / 225 GB | 2 TB | 25 Gbps | $2,190 | | NVIDIA RTX PRO 6000 Blackwell | 96 GB GDDR7 | 16 / 188 GB | 1.92 TB | 10 Gbps | $1,390 | | NVIDIA L40S | 48 GB GDDR6 | 16 / 128 GB | 1 TB | 10 Gbps | $890 | | NVIDIA GeForce RTX 5090 | 32 GB GDDR7 | 16 / 64 GB | 1 TB | 10 Gbps | $449 | A month is about 730 hours. Used around the clock, the H100 works out to about $3.00 per GPU-hour and the RTX 5090 to about $0.62. If you need a GPU for a few hours a week, hourly billing usually fits better. Our [monthly vs hourly comparison](https://overgpu.com/compare/monthly-vs-hourly-gpu-pricing) walks through the trade-off. ## When one GPU is enough Memory usually decides. As a rough rule, model weights take the parameter count times the bytes per parameter (2 in BF16, 1 in FP8, 0.5 in FP4). Add the KV cache and runtime overhead, then plan 10 to 30% headroom on top. - **Small models and image pipelines.** An 8B model needs about 16 GB in BF16, which leaves room for the KV cache on a 32 GB RTX 5090 or a 48 GB L40S. - **70B-class inference.** About 70 GB in FP8: tight on one H100, comfortable on an H200. In BF16 it is about 140 GB, so an H200 is tight and a B200 or B300 is comfortable. - **Fine-tuning.** LoRA and QLoRA need far less memory than full fine-tuning. Full fine-tuning with Adam takes roughly 16 to 20 bytes per parameter before activations, so even a 7B model needs about 112 to 140 GB. That points to a B200 or a B300. - **Rendering and generative media.** Large scenes, video models and big batches benefit from the 96 GB of the RTX PRO 6000. To run these numbers for your own model, precision and context length, use the [VRAM calculator](https://overgpu.com/tools/vram-calculator). ## Limits of a dedicated GPU VM A single-GPU plan is one GPU in one VM. Two plans are two separate machines that talk over the network, not over NVLink. When one model has to be split across GPUs, for example a model larger than the 288 GB of a B300 or a training job that needs tensor parallelism, choose an [8-GPU server](https://overgpu.com/products/8-gpu-servers) instead. It is bare metal, with NVLink between the GPUs on HGX systems and 50 TB of monthly bandwidth instead of 20 TB. The 8-GPU servers also cost slightly less per GPU. Eight single H100 plans add up to $17,520 a month, while an HGX H100 server is $16,900. Bare metal also gives you the whole machine: both CPUs, all system memory and every drive. ## Upgrades and ordering You are not locked into a size. The minimum term is one month, plans renew automatically from your account balance with an email reminder before each renewal, and you can cancel any month with no fee from your dashboard. The server then runs until the end of the paid month. To move up, say from an H100 to an H200 when your contexts grow, open a support ticket from your dashboard. Changes depend on stock, and we confirm the new configuration and price in writing before you pay; the new server is delivered in under 10 minutes after payment. Our engineers can advise on moving your data as part of [expert onboarding](https://overgpu.com/products/expert-onboarding), and the [upgrades and changes guide](https://overgpu.com/help/upgrades-and-changes) covers the details. A first order takes one form on the [deployment page](https://overgpu.com/deploy): choose your configuration, set up the OS image, optional preinstalled software and your login (root password or SSH key), create a free account or log in, and pay the first month from your balance, which you [top up in crypto](https://overgpu.com/pay-with-crypto). The server is delivered in under 10 minutes after payment, with its IP address and SSH details in your dashboard. ## Frequently asked questions ### Is the GPU shared with other customers? No. The whole physical GPU is passed through to your VM and assigned only to you, with no vGPU slicing, no MIG partitions and no time-sharing. ### What is the difference between a single-GPU plan and an 8-GPU server? A single-GPU plan is a dedicated VM with one GPU passed through and its own vCPU, RAM and NVMe. An 8-GPU server is a bare-metal machine with eight GPUs, NVLink between them on HGX systems, and 50 TB of monthly bandwidth instead of 20 TB. ### Can I run Docker containers on the GPU? Yes. The Ubuntu images include Docker and the NVIDIA Container Toolkit, and you have full root access to install anything allowed by our acceptable use policy. ### How much bandwidth is included? 20 TB of outbound traffic per month under fair use, with no egress fees. Inbound traffic, such as uploading datasets, is not counted. ### Can I move to a bigger GPU later? Yes, subject to stock: open a support ticket from your dashboard. We confirm the new configuration and price in writing before you pay, and the new server is delivered in under 10 minutes after payment. You can cancel the old plan, which runs until the end of its paid month. ## Related pages - [8-GPU server rental: bare-metal NVIDIA HGX and AMD Instinct](https://overgpu.com/products/8-gpu-servers): NVIDIA HGX B300, B200, H200, H100, AMD MI355X or 8x RTX PRO 6000. 50 TB bandwidth, monthly terms, delivered in under 10 minutes. - [Rent NVIDIA H200 and HGX H200 servers from $3,090 per month](https://overgpu.com/gpus/nvidia-h200): Rent a dedicated NVIDIA H200 (141 GB) for $3,090/month, about $4.23 per GPU-hour at 24/7, or an 8-GPU HGX H200 for $23,900. Flat price, paid in crypto. - [H100 vs H200: same Hopper compute, very different memory](https://overgpu.com/compare/h100-vs-h200): Same Hopper compute, 80 GB at 3.35 TB/s vs 141 GB at 4.8 TB/s. What fits on one GPU, the KV cache math and the monthly cost, compared. - [LLM VRAM calculator: will your model fit on a GPU?](https://overgpu.com/tools/vram-calculator): Pick Llama, Qwen, Mistral or gpt-oss, set the precision and context length, and see weights, KV cache and which GPUs fit. - [How to upgrade GPU server plans and make changes](https://overgpu.com/help/upgrades-and-changes): Upgrade GPU server plan mid-month with a prorated charge from your balance; smaller plans apply at your next renewal, subject to stock. See how to ask. - [How to choose a GPU server for AI](https://overgpu.com/blog/how-to-choose-a-gpu-server): A framework for memory, bandwidth, precision, interconnect, CPU and storage, software and billing, plus a checklist.