# GPU.ai

by GPU.ai · checked 2 Nov 2026 · updated 2 Nov 2026

Website: [gpu.ai](https://gpu.ai)

## Facts

- Offers: on-demand GPU cloud, reserved GPUs, a GPU marketplace, serverless inference, GPU clusters.
- GPU models: B200, H200, H100, A100, L40S, L40, L4, RTX PRO 6000, RTX 6000 Ada, RTX A6000, RTX 5090, RTX 4090, RTX 3090, other GPUs.
- Billing increment: billing by the second.
- Access: SSH, Jupyter, an API, a web console, own Docker images, templates.
- Sign-up needs: an e-mail address.
- Pay with: USDT, USDC.
- GPU.ai does not publish the iD check.

## Basics

| Attribute | Value | Applies to | Source |
| --- | --- | --- | --- |
| Offers | on-demand GPU cloud, reserved GPUs, a GPU marketplace, serverless inference, GPU clusters |  | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |
| Billing increment | billing by the second |  | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |
| Spot / interruptible | Yes |  | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |
| Access | SSH, Jupyter, an API, a web console, own Docker images, templates |  | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |
| Persistent storage | unknown |  |  |
| Free credits or trial | unknown |  |  |
| Refund policy | Technical issues (dead GPU, downtime, broken instance) are refunded for the affected usage after a support ticket; referral credit is non-withdrawable. |  | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |

## Fees and variants

|  | All variants | B200 | H200 SXM | H100 SXM | H100 PCIE | A100 | RTX PRO 6000 | RTX A6000 | L40S | L4 | RTX 4090 | RTX 5090 | RTX 2000 ADA | Source |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| Status |  | available | available | available | available | available | available | available | available | available | available | available | available |  |
| Price |  | Price not published | Price not published | Price not published | Price not published | Price not published | Price not published | Price not published | Price not published | Price not published | Price not published | Price not published | Price not published |  |
| Score |  | 8.3 (Entry variant) | 8.3 | 8.3 | 8.3 | 8.3 | 8.3 | 8.3 | 8.3 | 8.3 | 8.3 | 8.3 | 8.3 |  |
| **Provider** |  |  |  |  |  |  |  |  |  |  |  |  |  |  |
| GPU model |  | B200 | H200 | H100 | H100 | A100 | RTX PRO 6000 | RTX A6000 | L40S | L4 | RTX 4090 | RTX 5090 | other GPUs | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |
| GPUs |  | 1 | unknown | unknown | unknown | unknown | unknown | unknown | unknown | unknown | unknown | unknown | unknown | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |
| VRAM per GPU (GB) |  | 192 | 141 | 80 | 80 | 80 | 96 | 48 | 48 | 24 | 24 | 32 | 16 | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |
| **Fees** |  |  |  |  |  |  |  |  |  |  |  |  |  |  |
| Price per hour |  | 7.99 USD | 5.29 USD | 3.99 USD | 1.98 USD | 1.07 USD | 1.69 USD | 0.40 USD | 0.80 USD | 0.59 USD | 0.24 USD | 0.24 USD | 0.24 USD | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |
| Spot price per hour | unknown |  |  |  |  |  |  |  |  |  |  |  |  |  |
| Price per month | unknown |  |  |  |  |  |  |  |  |  |  |  |  |  |
| H100 per GPU-hour | 1.98 USD |  |  |  |  |  |  |  |  |  |  |  |  | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |
| A100 per GPU-hour | 1.07 USD |  |  |  |  |  |  |  |  |  |  |  |  | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |
| RTX 4090 per GPU-hour | 0.89 USD |  |  |  |  |  |  |  |  |  |  |  |  | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |

## Availability

| Attribute | Value | Applies to | Source |
| --- | --- | --- | --- |
| Customer countries | unknown |  |  |

| Country | Available |
| --- | --- |
| United States | unknown |
| Germany | unknown |
| United Kingdom | unknown |

## GPUs

| Attribute | Value | Applies to | Source |
| --- | --- | --- | --- |
| GPU models | B200, H200, H100, A100, L40S, L40, L4, RTX PRO 6000, RTX 6000 Ada, RTX A6000, RTX 5090, RTX 4090, RTX 3090, other GPUs |  | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |
| Max GPUs per instance | 8 |  | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |
| Multi-node clusters | Yes |  | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |
| GPU interconnect | unknown |  |  |
| GPU model | unknown |  |  |
| ↳ Plan B200 | B200 |  | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |
| ↳ Plan H200 SXM | H200 |  | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |
| ↳ Plan H100 SXM | H100 |  | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |
| ↳ Plan H100 PCIE | H100 |  | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |
| ↳ Plan A100 | A100 |  | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |
| ↳ Plan RTX PRO 6000 | RTX PRO 6000 |  | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |
| ↳ Plan RTX A6000 | RTX A6000 |  | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |
| ↳ Plan L40S | L40S |  | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |
| ↳ Plan L4 | L4 |  | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |
| ↳ Plan RTX 4090 | RTX 4090 |  | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |
| ↳ Plan RTX 5090 | RTX 5090 |  | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |
| ↳ Plan RTX 2000 ADA | other GPUs |  | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |
| GPUs | unknown |  |  |
| ↳ Plan B200 | 1 |  | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |
| VRAM per GPU (GB) | unknown |  |  |
| ↳ Plan B200 | 192 |  | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |
| ↳ Plan H200 SXM | 141 |  | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |
| ↳ Plan H100 SXM | 80 |  | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |
| ↳ Plan H100 PCIE | 80 |  | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |
| ↳ Plan A100 | 80 |  | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |
| ↳ Plan RTX PRO 6000 | 96 |  | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |
| ↳ Plan RTX A6000 | 48 |  | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |
| ↳ Plan L40S | 48 |  | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |
| ↳ Plan L4 | 24 |  | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |
| ↳ Plan RTX 4090 | 24 |  | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |
| ↳ Plan RTX 5090 | 32 |  | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |
| ↳ Plan RTX 2000 ADA | 16 |  | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |
| vCPU cores | unknown |  |  |
| RAM (GB) | unknown |  |  |
| Storage (GB) | unknown |  |  |

## Locations & hosts

| Attribute | Value | Applies to | Source |
| --- | --- | --- | --- |
| Data centre / host countries | unknown |  |  |
| Locations | 8 |  | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |
| Hosts | data centres, community hosts |  | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |
| Free egress | Yes |  | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |

## Privacy

| Attribute | Value | Applies to | Source |
| --- | --- | --- | --- |
| Sign-up needs | an e-mail address |  | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |
| ID check | not published |  | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |

## Rules

| Attribute | Value | Applies to | Source |
| --- | --- | --- | --- |
| Allowed uses | not applicable |  | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |

## Payment

| Attribute | Value | Applies to | Source |
| --- | --- | --- | --- |
| Pay with | USDT, USDC |  | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |
| Payment chains | Ethereum, Solana, Polygon, Arbitrum, Base |  | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |
| Other payment methods | card |  | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |
| Minimum deposit | 5.00 USD |  | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |

## Promotions

What GPU.ai offers itself: referral bonus, welcome bonus, running promotions and partner programme. Values of the provider, with source; no affiliate link in it.

- Referral bonus · You (referrer): 10 USD · Conditions: Paid as non-withdrawable GPU.ai credit after a 30-day hold, once the referred friend deposits $10 or more and incurs first real GPU or serverless usage. · invite link, code — gpu.ai

## Trust

| Attribute | Value | Applies to | Source |
| --- | --- | --- | --- |
| Uptime guarantee | unknown |  |  |
| Support channels | email |  | gpu.ai (checked 2 Nov 2026 · verified 2 Nov 2026) |

## Company and trust

### Company

- Headquarters: United Arab Emirates (Source: gpu.ai)
- Legal form: Private Company (Source: gpu.ai)
- Register: DIFC Registrar of Companies (Source: gpu.ai)
- Registered address: Office IH-00-01-02-OF-01, Level 2, Innovation One, Dubai International Financial Centre, Dubai, United Arab Emirates (Source: gpu.ai)

### Transparency

- Transparency 78 %
- 1 of 11 key values not published
- Imprint / company details: yes
- Public fee schedule: yes
- Terms and conditions: no
- Terms with version or date: unknown
- Sources last checked 2026-11-02

## How the score is made up

8.3 of 10 · very good. Weighted sum of the criteria, each 0 to 10 from proven values. A criterion without data is not scored and counts with the category median, not as 0.

| Criterion | Weight | Part (0–10) | Points |
| --- | --- | --- | --- |
| Trust | 25 % | 8.5 | 2.1 |
| Features | 20 % | 10.0 | 2.0 |
| Anonymity | 20 % | 6.7 | 1.3 |
| GPU choice | 15 % | 10.0 | 1.5 |
| Costs | 10 % | 9.3 | 0.9 |
| Locations | 5 % | not scored | counts with the median (0.2) |
| Community | 5 % | not scored | counts neutral (0.3) |

## Changelog

- 2 Nov 2026: Score changed from 80.45 to 82.94.
- 2 Nov 2026: New: Referral bonus · You (referrer).
- 2 Nov 2026: Headquarters: United Arab Emirates.
- 2 Nov 2026: Score changed from 80.44 to 80.45.
- 2 Nov 2026: Score changed from 80.42 to 80.44.
- 2 Nov 2026: Score changed from 80.41 to 80.42.
- 2 Nov 2026: Score changed from 80.38 to 80.41.
- 2 Nov 2026: Score changed from 80.64 to 80.38.
- 2 Nov 2026: Score changed from 78.78 to 80.64.

Found an error or a missing value? Report it with a source: https://cryptogpuss.com/en/gpuai/report-an-error/

---

- **Canonical URL:** https://cryptogpuss.com/en/gpuai/
- **As of:** 2 Nov 2026
- **Data licence:** CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/)

Please cite: Source: cryptogpuss.com (cryptovss), as of 2 Nov 2026, https://cryptogpuss.com/en/gpuai/