NVIDIA
Rent an RTX 5090 server.
32 GB of GDDR7, $134.99 a month .
A dedicated
32 GB GDDR7 ,1,792 GB/s - One or two cards
per server - Ready in
1 to 24 hours No KYC , crypto only
01Plans & pricing
One or two RTX 5090s per server .
Every card is yours alone, with NVMe storage and an unmetered
| GPU | VRAM | CPU | RAM | Storage | Locations | Price | Order |
|---|---|---|---|---|---|---|---|
   |
|
$134.99/mo
Cheapest competitor |
Deploy |
||||
|
|
|
$558.99/mo
Cheapest competitor |
Deploy |
No GPU plan in this jurisdiction yet: or compare them.
RTX 4090 , 5090, L40S, A100, H100- CUDA drivers preinstalled
- Single and
multi-GPU nodes No KYC · BTC, ETH, XMR, USDT, SOL
Want the value pick? The
02Specifications
The RTX 5090 , in numbers.
NVIDIA’s specifications for the card, and what comes with it in our servers.
| Specification | ||
|---|---|---|
| Architecture | NVIDIA Blackwell | NVIDIA Blackwell |
| CUDA cores | 21,760 | |
| GPU memory | ||
| Memory bandwidth | ||
| Tensor cores | 5th generation, with FP4 and FP8 | 5th generation, with FP4 and FP8 |
| FP32 compute | 104.8 TFLOPS | |
| Between the cards | Single card | PCIe, no NVLink |
| CPU, RAM, storage | ||
| Price | $134.99/mo, about $0.185/h | $558.99/mo, about $0.766/h |
03Sizing
What runs on 32 and 64 GB .
The highest precision at which a dense language model fits with an
| Model size | ||
|---|---|---|
| 7–8B | ||
| 13–14B | ||
| 32B | ||
| 70B | Does not fit | |
| 123B | Does not fit | Does not fit |
Two cards split one model with vLLM or llama.cpp. The full method: how much VRAM does an LLM need?
04Monthly or hourly
When a month beats paying by the hour.
At $134.99
| If the hourly rate is | Our month equals | |
|---|---|---|
| $0.70 per | 26% of the month | |
| $1.00 per | 18% of the month | |
| $1.50 per | 12% of the month |
Hourly rates are examples, not quotes from any provider. We bill monthly, quarterly (−5%) or yearly (−12%), from your balance.
05Use cases
What an RTX 5090 is best at.
More memory and much more bandwidth than the 4090: faster tokens and longer contexts.
Fast inference
14B models at
8-bit with long contexts, or 32B models at4-bit , answering noticeably faster than on a 4090.70B at
4-bit Two RTX 5090s hold a
4-bit 70B model across64 GB , split with vLLM or llama.cpp.Image and video
Flux and video models that do not fit comfortably in
24 GB .Rendering and mining
Blender Cycles renders, and mining, which is allowed on GPU servers.
06Jurisdictions
Where RTX 5090 servers run.
The plans above show where each server is available; prices are the same everywhere.
-
MoldovaChișinăuBest value
Outside the EU
· no DSA London~33 ms New York ~113 ms Singapore~129 ms -
NetherlandsAmsterdamNetwork hub
EU member
· DSA applies London~7 ms New York ~87 ms Singapore~154 ms -
IcelandReykjavík
Free-speech haven EEA· outside the EU· no DSA London~29 ms New York ~63 ms Singapore~169 ms
Latency figures are estimates from distance, not measurements. Every jurisdiction has the same prices. How we estimate latency
07Offshore, in writing
The rules, before you pay.
What we promise is written into our policies, not just our marketing.
US DMCA noticesNot actionedAnswered with our policy, never enforced. Only a local
DMCA policycourt order , or a valid EU notice in EU locations, can require action.IdentityEmail only
No name, address, phone or ID document, ever. A private or disposable address is fine.
Privacy policyPayment5 cryptocurrencies
Bitcoin, Ethereum, Monero, USDT and Solana, paid
Crypto paymentson-chain from any wallet to your balance. No card processor, no chargebacks.TransparencySigned canary
A
Warrant canaryPGP-signed warrant canary every quarter and a public count of every request we receive.
08FAQ
RTX 5090 servers, answered.
Another question? The full FAQ answers what people ask before an order: payments, privacy, complaints and support.
Read the full FAQRTX 5090 or RTX 4090 ?
The 5090 has
Does PyTorch support the RTX 5090 ?
Yes, with a recent CUDA build: Blackwell cards need newer CUDA versions than older GPUs. The GPU images guide gives the exact install commands.
Are the two cards linked by NVLink?
No. The two RTX 5090s talk over PCIe. Splitting a model for inference works well; for
Can I rent it by the hour?
No. Billing is monthly, quarterly (−5%) or yearly (−12%), paid from your balance. At about $0.185 an hour around the clock,
How fast is delivery?
Between 1 and
Can I get a refund?
GPU servers are not refundable once delivered, because the hardware is set aside for you. Check the specifications above before you order.
Blackwell speed, offshore and private.
- 1Pick one card or two
- 2Pay in crypto,
no ID asked - 3SSH in within
1 to 24 hours