# Dedicated GPU servers, *paid in crypto.*

> Dedicated NVIDIA GPU servers by the month: H100, A100, RTX 4090, H200, B200. Paid in crypto, no identity check. Full root, unmetered bandwidth, one tenant per card.

gpuserver.io rents dedicated NVIDIA GPU servers by the month, paid in cryptocurrency and opened without a single identity check. One tenant per card, full root access and an out-of-band console — in six data centres, from $125 a month.

- Bitcoin, Monero, USDT & 5 more
- No KYC, no documents, no card

[Browse 41 configurations](https://gpuserver.io/#catalog) [Crypto & no KYC, explained](https://gpuserver.io/#payment)

What will it run?

Cheapest machine that fits

NVIDIA L4

24 GB GDDR6 · Intel Xeon Silver 4410Y · 128 GB RAM

**$125** / month

LoRA and QLoRA on 7B to 13B · needs 24 GB per card

[Show it in the catalogue](https://gpuserver.io/#catalog)

- [**< 5 min** from payment to root access](https://gpuserver.io/docs#delivery)
- [**99.9%** uptime, or we extend your term](https://gpuserver.io/legal/sla)
- [**No KYC** no documents, at any spend](https://gpuserver.io/no-kyc-gpu-server)
- [**$0** setup fee on every machine](https://gpuserver.io/#catalog)

## Every machine we rent, on one line each.

Refine

GPU12

GPUs per server5

Memory per card7

System RAM4

Interconnect2

**41** of 41 configurations · every one deployable in all 6 data centres · monthly billing, cancel any time

- [L4](https://gpuserver.io/gpu/nvidia-l4) GS-L4 · Ada Lovelace · PCIe 5.0   **24 GB**   **128 GB** Intel Xeon Silver 4410Y **2 × 2 TB NVMe** 12 c / 24 t **1 Gbit/s** PCIe 5.0  **$125** billed monthly  [Configure](https://gpuserver.io/configure?c=l4-x1)
- [RTX 4090](https://gpuserver.io/gpu/rtx-4090) GS-RTX4090 · Ada Lovelace · PCIe 5.0   **24 GB**   **128 GB** AMD Ryzen 9 7950X **2 × 2 TB NVMe** 16 c / 32 t **1 Gbit/s** PCIe 5.0  **$193** billed monthly  [Configure](https://gpuserver.io/configure?c=rtx4090-x1)
- [2 × L4](https://gpuserver.io/gpu/nvidia-l4) GN-L4×2 · Ada Lovelace · PCIe 5.0 ×16   **48 GB**   **256 GB** Intel Xeon Gold 5416S **2 × 3.84 TB NVMe** 16 c / 32 t **2 Gbit/s** PCIe 5.0 ×16  **$285** billed monthly  [Configure](https://gpuserver.io/configure?c=l4-x2)
- [RTX A6000](https://gpuserver.io/gpu/rtx-a6000) GS-A6000 · Ampere · PCIe 5.0   **48 GB**   **128 GB** AMD Ryzen 9 7950X **2 × 2 TB NVMe** 16 c / 32 t **1 Gbit/s** PCIe 5.0  **$286** billed monthly  [Configure](https://gpuserver.io/configure?c=a6000-x1)
- [RTX 5090](https://gpuserver.io/gpu/rtx-5090) GS-RTX5090 · Blackwell · PCIe 5.0   **32 GB**   **128 GB** AMD Ryzen 9 7950X **2 × 2 TB NVMe** 16 c / 32 t **1 Gbit/s** PCIe 5.0  **$335** billed monthly  [Configure](https://gpuserver.io/configure?c=rtx5090-x1)
- [A100 PCIe](https://gpuserver.io/gpu/a100-40gb) GS-A10040 · Ampere · PCIe 5.0   **40 GB**   **128 GB** Intel Xeon Silver 4410Y **2 × 2 TB NVMe** 12 c / 24 t **1 Gbit/s** PCIe 5.0  **$392** billed monthly  [Configure](https://gpuserver.io/configure?c=a100-40-x1)
- [2 × RTX 4090](https://gpuserver.io/gpu/rtx-4090) GN-RTX4090×2 · Ada Lovelace · PCIe 5.0 ×16   **48 GB**   **256 GB** AMD Threadripper 7960X **2 × 3.84 TB NVMe** 24 c / 48 t **2 Gbit/s** PCIe 5.0 ×16  **$416** billed monthly  [Configure](https://gpuserver.io/configure?c=rtx4090-x2)
- [4 × L4](https://gpuserver.io/gpu/nvidia-l4) GN-L4×4 · Ada Lovelace · PCIe 5.0 ×16   **96 GB**   **512 GB** 2 × Intel Xeon Gold 6438Y+ **4 × 3.84 TB NVMe** 64 c **10 Gbit/s** PCIe 5.0 ×16  **$564** billed monthly  [Configure](https://gpuserver.io/configure?c=l4-x4)
- [2 × RTX A6000](https://gpuserver.io/gpu/rtx-a6000) GN-A6000×2 · Ampere · PCIe 5.0 ×16   **96 GB**   **256 GB** AMD Threadripper 7960X **2 × 3.84 TB NVMe** 24 c / 48 t **2 Gbit/s** PCIe 5.0 ×16  **$597** billed monthly  [Configure](https://gpuserver.io/configure?c=a6000-x2)
- [2 × RTX 5090](https://gpuserver.io/gpu/rtx-5090) GN-RTX5090×2 · Blackwell · PCIe 5.0 ×16   **64 GB**   **256 GB** AMD Threadripper 7960X **2 × 3.84 TB NVMe** 24 c / 48 t **2 Gbit/s** PCIe 5.0 ×16  **$692** billed monthly  [Configure](https://gpuserver.io/configure?c=rtx5090-x2)
- [L40S](https://gpuserver.io/gpu/l40s) GS-L40S · Ada Lovelace · PCIe 5.0   **48 GB**   **128 GB** Intel Xeon Silver 4410Y **2 × 2 TB NVMe** 12 c / 24 t **1 Gbit/s** PCIe 5.0  **$714** billed monthly  [Configure](https://gpuserver.io/configure?c=l40s-x1)
- [2 × A100 PCIe](https://gpuserver.io/gpu/a100-40gb) GN-A10040×2 · Ampere · PCIe 5.0 ×16   **80 GB**   **256 GB** Intel Xeon Gold 5416S **2 × 3.84 TB NVMe** 16 c / 32 t **2 Gbit/s** PCIe 5.0 ×16  **$802** billed monthly  [Configure](https://gpuserver.io/configure?c=a100-40-x2)
- [4 × RTX 4090](https://gpuserver.io/gpu/rtx-4090) GN-RTX4090×4 · Ada Lovelace · PCIe 5.0 ×16   **96 GB**   **512 GB** AMD Threadripper 7970X **4 × 3.84 TB NVMe** 32 c / 64 t **10 Gbit/s** PCIe 5.0 ×16  **$814** billed monthly  [Configure](https://gpuserver.io/configure?c=rtx4090-x4)
- [A100 PCIe](https://gpuserver.io/gpu/a100-80gb) GS-A10080 · Ampere · PCIe 5.0   **80 GB**   **128 GB** Intel Xeon Silver 4410Y **2 × 2 TB NVMe** 12 c / 24 t **1 Gbit/s** PCIe 5.0  **$1,091** billed monthly  [Configure](https://gpuserver.io/configure?c=a100-80-x1)
- [8 × L4](https://gpuserver.io/gpu/nvidia-l4) GN-L4×8 · Ada Lovelace · PCIe 5.0 ×16   **192 GB**   **1 TB** 2 × Intel Xeon Platinum 8462Y+ **8 × 3.84 TB NVMe** 64 c **25 Gbit/s** PCIe 5.0 ×16  **$1,106** billed monthly  [Configure](https://gpuserver.io/configure?c=l4-x8)
- [4 × RTX A6000](https://gpuserver.io/gpu/rtx-a6000) GN-A6000×4 · Ampere · PCIe 5.0 ×16   **192 GB**   **512 GB** AMD Threadripper 7970X **4 × 3.84 TB NVMe** 32 c / 64 t **10 Gbit/s** PCIe 5.0 ×16  **$1,163** billed monthly  [Configure](https://gpuserver.io/configure?c=a6000-x4)
- [4 × RTX 5090](https://gpuserver.io/gpu/rtx-5090) GN-RTX5090×4 · Blackwell · PCIe 5.0 ×16   **128 GB**   **512 GB** AMD Threadripper 7970X **4 × 3.84 TB NVMe** 32 c / 64 t **10 Gbit/s** PCIe 5.0 ×16  **$1,345** billed monthly  [Configure](https://gpuserver.io/configure?c=rtx5090-x4)
- [10 × L4](https://gpuserver.io/gpu/nvidia-l4) GN-L4×10 · Ada Lovelace · PCIe 5.0 ×16   **240 GB**   **1 TB** 2 × Intel Xeon Platinum 8462Y+ **8 × 7.68 TB NVMe** 64 c **25 Gbit/s** PCIe 5.0 ×16  **$1,371** billed monthly  [Configure](https://gpuserver.io/configure?c=l4-x10)
- [2 × L40S](https://gpuserver.io/gpu/l40s) GN-L40S×2 · Ada Lovelace · PCIe 5.0 ×16   **96 GB**   **256 GB** Intel Xeon Gold 5416S **2 × 3.84 TB NVMe** 16 c / 32 t **2 Gbit/s** PCIe 5.0 ×16  **$1,427** billed monthly  [Configure](https://gpuserver.io/configure?c=l40s-x2)
- [H100 PCIe](https://gpuserver.io/gpu/h100-pcie) GS-H100PCIE · Hopper · PCIe 5.0   **80 GB**   **128 GB** Intel Xeon Silver 4410Y **2 × 2 TB NVMe** 12 c / 24 t **1 Gbit/s** PCIe 5.0  **$1,469** billed monthly  [Configure](https://gpuserver.io/configure?c=h100-pcie-x1)
- [4 × A100 PCIe](https://gpuserver.io/gpu/a100-40gb) GN-A10040×4 · Ampere · PCIe 5.0 ×16   **160 GB**   **512 GB** 2 × Intel Xeon Gold 6438Y+ **4 × 3.84 TB NVMe** 64 c **10 Gbit/s** PCIe 5.0 ×16  **$1,556** billed monthly  [Configure](https://gpuserver.io/configure?c=a100-40-x4)
- [8 × RTX 4090](https://gpuserver.io/gpu/rtx-4090) GN-RTX4090×8 · Ada Lovelace · PCIe 5.0 ×16   **192 GB**   **1 TB** 2 × Intel Xeon Gold 6438Y+ **8 × 3.84 TB NVMe** 64 c **25 Gbit/s** PCIe 5.0 ×16  **$1,579** billed monthly  [Configure](https://gpuserver.io/configure?c=rtx4090-x8)
- [2 × A100 PCIe](https://gpuserver.io/gpu/a100-80gb) GN-A10080×2 · Ampere · PCIe 5.0 ×16   **160 GB**   **256 GB** Intel Xeon Gold 5416S **2 × 3.84 TB NVMe** 16 c / 32 t **2 Gbit/s** PCIe 5.0 ×16  **$2,159** billed monthly  [Configure](https://gpuserver.io/configure?c=a100-80-x2)
- [8 × RTX A6000](https://gpuserver.io/gpu/rtx-a6000) GN-A6000×8 · Ampere · PCIe 5.0 ×16   **384 GB**   **1 TB** 2 × Intel Xeon Gold 6438Y+ **8 × 3.84 TB NVMe** 64 c **25 Gbit/s** PCIe 5.0 ×16  **$2,239** billed monthly  [Configure](https://gpuserver.io/configure?c=a6000-x8)
- [8 × RTX 5090](https://gpuserver.io/gpu/rtx-5090) GN-RTX5090×8 · Blackwell · PCIe 5.0 ×16   **256 GB**   **1 TB** 2 × Intel Xeon Gold 6438Y+ **8 × 3.84 TB NVMe** 64 c **25 Gbit/s** PCIe 5.0 ×16  **$2,584** billed monthly  [Configure](https://gpuserver.io/configure?c=rtx5090-x8)
- [4 × L40S](https://gpuserver.io/gpu/l40s) GN-L40S×4 · Ada Lovelace · PCIe 5.0 ×16   **192 GB**   **512 GB** 2 × Intel Xeon Gold 6438Y+ **4 × 3.84 TB NVMe** 64 c **10 Gbit/s** PCIe 5.0 ×16  **$2,754** billed monthly  [Configure](https://gpuserver.io/configure?c=l40s-x4)
- [2 × H100 PCIe](https://gpuserver.io/gpu/h100-pcie) GN-H100PCIE×2 · Hopper · PCIe 5.0 ×16   **160 GB**   **256 GB** Intel Xeon Gold 5416S **2 × 3.84 TB NVMe** 16 c / 32 t **2 Gbit/s** PCIe 5.0 ×16  **$2,892** billed monthly  [Configure](https://gpuserver.io/configure?c=h100-pcie-x2)
- [8 × A100 PCIe](https://gpuserver.io/gpu/a100-40gb) GN-A10040×8 · Ampere · PCIe 5.0 ×16   **320 GB**   **1 TB** 2 × Intel Xeon Platinum 8462Y+ **8 × 3.84 TB NVMe** 64 c **25 Gbit/s** PCIe 5.0 ×16  **$2,983** billed monthly  [Configure](https://gpuserver.io/configure?c=a100-40-x8)
- [4 × A100 PCIe](https://gpuserver.io/gpu/a100-80gb) GN-A10080×4 · Ampere · PCIe 5.0 ×16   **320 GB**   **512 GB** 2 × Intel Xeon Gold 6438Y+ **4 × 3.84 TB NVMe** 64 c **10 Gbit/s** PCIe 5.0 ×16  **$4,157** billed monthly  [Configure](https://gpuserver.io/configure?c=a100-80-x4)
- [4 × A100 SXM4](https://gpuserver.io/gpu/a100-sxm4) GN-A100SXM×4 · Ampere · NVLink 3 · 600 GB/s   **320 GB**   **512 GB** 2 × Intel Xeon Gold 6438Y+ **4 × 3.84 TB NVMe** 64 c **10 Gbit/s** NVLink 3 · 600 GB/s  **$4,395** billed monthly  [Configure](https://gpuserver.io/configure?c=a100-sxm-x4)
- [8 × L40S](https://gpuserver.io/gpu/l40s) GN-L40S×8 · Ada Lovelace · PCIe 5.0 ×16   **384 GB**   **1 TB** 2 × Intel Xeon Platinum 8462Y+ **8 × 3.84 TB NVMe** 64 c **25 Gbit/s** PCIe 5.0 ×16  **$5,251** billed monthly  [Configure](https://gpuserver.io/configure?c=l40s-x8)
- [4 × H100 PCIe](https://gpuserver.io/gpu/h100-pcie) GN-H100PCIE×4 · Hopper · PCIe 5.0 ×16   **320 GB**   **512 GB** 2 × Intel Xeon Gold 6438Y+ **4 × 3.84 TB NVMe** 64 c **10 Gbit/s** PCIe 5.0 ×16  **$5,564** billed monthly  [Configure](https://gpuserver.io/configure?c=h100-pcie-x4)
- [4 × H100 SXM5](https://gpuserver.io/gpu/h100-sxm5) GN-H100SXM×4 · Hopper · NVLink 4 · 900 GB/s   **320 GB**   **512 GB** 2 × Intel Xeon Gold 6438Y+ **4 × 3.84 TB NVMe** 64 c **10 Gbit/s** NVLink 4 · 900 GB/s  **$6,061** billed monthly  [Configure](https://gpuserver.io/configure?c=h100-sxm-x4)
- [8 × A100 PCIe](https://gpuserver.io/gpu/a100-80gb) GN-A10080×8 · Ampere · PCIe 5.0 ×16   **640 GB**   **1 TB** 2 × Intel Xeon Platinum 8462Y+ **8 × 3.84 TB NVMe** 64 c **25 Gbit/s** PCIe 5.0 ×16  **$7,906** billed monthly  [Configure](https://gpuserver.io/configure?c=a100-80-x8)
- [8 × A100 SXM4](https://gpuserver.io/gpu/a100-sxm4) GN-A100SXM×8 · Ampere · NVLink 3 · 600 GB/s   **640 GB**   **1 TB** 2 × Intel Xeon Platinum 8462Y+ **8 × 3.84 TB NVMe** 64 c **25 Gbit/s** NVLink 3 · 600 GB/s  **$8,355** billed monthly  [Configure](https://gpuserver.io/configure?c=a100-sxm-x8)
- [4 × H200 SXM5](https://gpuserver.io/gpu/h200) GN-H200×4 · Hopper · NVLink 4 · 900 GB/s   **564 GB**   **512 GB** 2 × Intel Xeon Gold 6438Y+ **4 × 3.84 TB NVMe** 64 c **10 Gbit/s** NVLink 4 · 900 GB/s  **$8,405** billed monthly  [Configure](https://gpuserver.io/configure?c=h200-x4)
- [8 × H100 PCIe](https://gpuserver.io/gpu/h100-pcie) GN-H100PCIE×8 · Hopper · PCIe 5.0 ×16   **640 GB**   **1 TB** 2 × Intel Xeon Platinum 8462Y+ **8 × 3.84 TB NVMe** 64 c **25 Gbit/s** PCIe 5.0 ×16  **$10,568** billed monthly  [Configure](https://gpuserver.io/configure?c=h100-pcie-x8)
- [8 × H100 SXM5](https://gpuserver.io/gpu/h100-sxm5) GN-H100SXM×8 · Hopper · NVLink 4 · 900 GB/s   **640 GB**   **1 TB** 2 × Intel Xeon Platinum 8462Y+ **8 × 3.84 TB NVMe** 64 c **25 Gbit/s** NVLink 4 · 900 GB/s  **$11,509** billed monthly  [Configure](https://gpuserver.io/configure?c=h100-sxm-x8)
- [4 × B200 SXM6](https://gpuserver.io/gpu/b200) GN-B200×4 · Blackwell · NVLink 5 · 1.8 TB/s   **720 GB**   **512 GB** 2 × Intel Xeon Gold 6438Y+ **4 × 3.84 TB NVMe** 64 c **10 Gbit/s** NVLink 5 · 1.8 TB/s  **$11,790** billed monthly  [Configure](https://gpuserver.io/configure?c=b200-x4)
- [8 × H200 SXM5](https://gpuserver.io/gpu/h200) GN-H200×8 · Hopper · NVLink 4 · 900 GB/s   **1128 GB**   **1 TB** 2 × Intel Xeon Platinum 8462Y+ **8 × 3.84 TB NVMe** 64 c **25 Gbit/s** NVLink 4 · 900 GB/s  **$15,945** billed monthly  [Configure](https://gpuserver.io/configure?c=h200-x8)
- [8 × B200 SXM6](https://gpuserver.io/gpu/b200) GN-B200×8 · Blackwell · NVLink 5 · 1.8 TB/s   **1440 GB**   **1 TB** 2 × Intel Xeon Platinum 8462Y+ **8 × 3.84 TB NVMe** 64 c **25 Gbit/s** NVLink 5 · 1.8 TB/s  **$22,351** billed monthly  [Configure](https://gpuserver.io/configure?c=b200-x8)

More RAM, more disk, extra addresses and private networking are monthly options in [the configurator](https://gpuserver.io/configure), on any machine above. If none of these hold your model, [the sizing guide](https://gpuserver.io/guides/vram-sizing) shows the arithmetic that says which one does — or that none of them do.

What you get

## On every machine, without asking.

Six things that are not options, not a tier and not a separate line on the invoice — and they are identical on the cheapest single card and on a ten-GPU node.

- ### No identity check

  No document, no selfie, no phone number, no company registration.

  Nothing verified · at any spend
- ### The whole card

  The GPU is passed through to one machine, and that machine is yours.

  No MIG · no vGPU · no hypervisor
- ### Root and IPMI

  Full root from the first minute, plus remote power and virtual media.

  Out-of-band on its own VLAN
- ### Unmetered bandwidth

  No included volume, no overage tier, nothing to watch on a graph.

  1 to 25 Gbit/s · both directions
- ### No setup or reinstall fee

  Reimage as often as you like, to any operating system we offer.

  $0 · reinstalls unlimited
- ### Hardware swapped, fast

  From spares held in the same suite, at any hour, disks left in place.

  Four-hour target · 24/7

Every one of these is a paid option, a higher tier or an outright refusal somewhere else. Here they are the machine.

Payment & privacy

## Paid in crypto. Opened without an identity check.

The same terms on every account, from the cheapest single card to an eight-GPU node.

What we never ask you for

- An ID document
- A selfie
- A phone number
- A company registration
- A credit card
- A verification tier
- Proof of address

Opening an account takes an email, a password and a billing address for the invoice — typed by you, printed as typed, checked against nothing.

From picking a machine to root access

1. 01 Pick the machine A configuration and a data centre. No account needed to see a price.
2. 02 Pay in crypto One address, one payment, one confirmation. No card, no bank.
3. 03 It provisions itself Your image is written and the driver stack verified, automatically.
4. 04 Root arrives SSH, IPMI and the network block — under 5 minutes after payment.

Send less than the invoice and it stays open for the difference — nothing is lost. Send more and the surplus becomes extra days on the same machine.

Accepted today

- **Bitcoin** BTC · Bitcoin
- **Monero** XMR · Monero
- **Ethereum (ERC-20)** ETH · ERC-20
- **Tether (ERC-20)** USDT · ERC-20
- **Tether (TRC-20)** USDTTRC · TRC-20
- **Litecoin** LTC · Litecoin
- **TRON** TRX · TRON
- **Solana** SOL · Solana

No bank transfer, and that is deliberate: a transfer would identify you and defeat the point of everything above.

Where it runs, and what we owe you

## Six data centres. One commitment.

Our own racks, our own hardware, an engineer on site around the clock.

- AMS Amsterdam Netherlands · Tier III 100% wind
- STO Stockholm Sweden · Tier III 100% hydro
- ZRH Zürich Switzerland · Tier IV 96% carbon-free
- REY Reykjavík Iceland · Tier III 100% geothermal + hydro
- YUL Montréal Canada · Tier III 99% hydro
- SIN Singapore Singapore · Tier III Grid + REC offset

- **99.9%** uptime per machine, per month — miss it and we add the lost time to your term
- **< 30 min** to a first human reply, at any hour, from an engineer who can reach the rack
- **4 h** target from a confirmed hardware fault to a working machine, using on-site spares
- **72 h** your disks are kept untouched after a term ends, so a missed date costs nothing

Questions

## The things people ask before they order.

Is the GPU really dedicated?

Yes, and it is worth being precise about what that means. The card is passed through to a single physical machine that only you have an account on. There is no MIG partitioning, no vGPU layer and no other tenant scheduled on the same silicon. On multi-GPU nodes you get the whole node, including the NVLink or PCIe fabric between the cards.

How quickly is a server actually delivered?

In under 5 minutes. Every configuration on the list is available in all six data centres, racked and waiting with an image ready to write, so a payment confirming at three in the morning is a machine at three in the morning — provisioning is automated end to end and nobody has to wake up for it.

What happens if I need to cancel?

A monthly server can be cancelled before its renewal date and simply stops at the end of the paid month. There is no cancellation fee and no notice period. Longer terms are paid up front and are not refunded pro rata, which is the trade you make for the discount.

How do I pay?

In cryptocurrency, and only in cryptocurrency: Bitcoin, Monero, Ethereum (ERC-20), Tether (ERC-20), Tether (TRC-20), Litecoin, TRON and Solana. Each invoice gets its own address, the amount is fixed at the rate shown when the invoice is issued, and the machine is provisioned on the first confirmation. There is no card on file, no direct debit, and nothing to cancel with a bank the day you leave. We do not take bank transfers, precisely because a transfer would identify you and defeat the point.

Is my data encrypted or inspected?

We do not inspect machine disks and we do not keep logs of what runs inside them. Encryption at rest is yours to configure, because a key we hold is a key we can be compelled to hand over. When a server is decommissioned or reinstalled its disks are erased with a full overwrite before the machine is racked again.

Do you really not ask for identity documents?

Really. There is no document upload, no selfie, no phone number and no company registration — and nothing you type is checked against anything. Opening an account takes an email address, a password and a billing address, because an invoice has to carry one; we print it exactly as you type it and we never verify it. There is no verification tier either: an 8-GPU node is opened on exactly the same terms as the cheapest single card. What we necessarily know is what the service itself creates — your email, your invoices, and the IP addresses assigned to your machine.

What is the uptime commitment?

99.9% per calendar month on network and power, measured per machine. When we miss it we do not credit an account — there is no balance here to credit — we extend your term by the time you lost, automatically. You do not have to notice it, calculate it or ask for it.

Who is behind gpuserver.io?

A small independent team that owns the hardware it rents. The machines are ours, they sit in leased space in the six data centres listed on this page, and nothing is resold from another provider — which is why the specification on a row is the specification you get.

## Tell us the workload. We will tell you the machine.

Every configuration on the list is available, in all 6 data centres, and running under 5 minutes after your payment confirms — at any hour. Not sure which card fits? Describe the job and we will size it honestly.

[Browse the catalogue](https://gpuserver.io/#catalog) [Build it in the configurator](https://gpuserver.io/configure)

No sales call, no account needed to get a price, and no setup fee on any configuration.

---

Source: https://gpuserver.io/. This file is generated from the same data as the website; if a figure here differs from a page, the page is authoritative and this file is stale — the canonical source is https://gpuserver.io/.