GPU dedicated server hosting
Two ways to get a server with a GPU: GPU VPS with the whole card passed through, ready to order, or a dedicated GPU server built to your spec on request.
Best-fit dedicated plans
These are the closest stock plans in our catalogue for GPU dedicated server workloads. Fully customizable — port, storage, RAM, DDoS profile.
GPU VPS · RTX PRO 4500
16 vCPU · 128 GB DDR5 · 1 TB NVMe Gen5 · RTX PRO 4500 (32 GB VRAM) full passthrough
GPU VPS · RTX PRO 6000
32 vCPU · 256 GB DDR5 · 2 TB NVMe Gen5 · RTX PRO 6000 (96 GB VRAM) full passthrough
Dedicated GPU server
Built to your spec: GPU count, CPU, RAM, storage and port agreed in a quote
About GPU dedicated server
If you search for a GPU dedicated server, you usually need two things: a GPU that nobody else shares, and a network that does not meter the data you move. Our GPU VPS passes the entire card into your virtual machine, and dedicated GPU servers are available on request.
Ready to order: GPU VPS with an NVIDIA RTX PRO 4500 Blackwell (32 GB VRAM) at $550/mo or an RTX PRO 6000 Blackwell (96 GB VRAM) at $1,200/mo, with full GPU passthrough. Details and the order form are on the [GPU VPS page](/gpu). Typical uses: LLM inference, Stable Diffusion and Flux, NVENC video transcoding and 3D rendering.
Dedicated GPU servers (bare metal, your own chassis) are not a catalogue item. Tell us the GPU count, VRAM, CPU, RAM, storage and port you need through the [contact form](/contact) and we will quote it. We do not list GPU models for dedicated servers here until a quote is confirmed.
Pair a GPU server with an unmetered port and Vault storage for datasets and media. For AI workloads see [AI inference servers](/ai-inference-server) and [deep learning servers](/deep-learning-server); for video, see [video transcoding servers](/video-transcoding-server).
Frequently asked questions
Do you offer dedicated GPU servers?
Yes, on request. Send us the GPU, CPU, RAM, storage and port requirements through the contact form and we will quote a build.
What is the difference between GPU VPS and a dedicated GPU server?
GPU VPS is a virtual machine that receives the whole GPU through passthrough and is ordered online. A dedicated GPU server is a physical server built for you, quoted individually.
Which GPUs are available right now?
GPU VPS: NVIDIA RTX PRO 4500 Blackwell (32 GB) and RTX PRO 6000 Blackwell (96 GB). Other GPUs for dedicated servers are decided per quote.
How much does it cost?
GPU VPS is $550/mo (RTX PRO 4500) and $1,200/mo (RTX PRO 6000). Dedicated GPU servers are priced per build.
Can I use it for AI inference and training?
Yes. Full passthrough with 32 GB or 96 GB VRAM suits LLM inference, diffusion models and fine-tuning. Larger training jobs may need a multi-GPU build, quoted on request.
Is bandwidth metered?
GPU VPS uses a 10 Gbit/s unmetered port as listed on the GPU page. Dedicated GPU builds can be paired with higher unmetered ports, quoted per project.