At IteraGPU, we help you fund an ML campaign with a rental amount known before ordering. Start at USD 23.57 for 1 lot of NVIDIA RTX A5000 24GB, i.e. 1 GPUs with 24 GB per card for 3 days, or choose a 7- or 30-day package. Pay directly in crypto, with no forced top-up and no KYC procedure or ID document required. An account with first name, last name, email and password is required to order.
Account required: first name, last name, email and password. No KYC procedure and no ID document.
We display the 45 packages in USD, by configuration and 3-, 7- or 30-day period. For a campaign whose working window you know, choose this full amount: you can compare eligible configurations before placing your order.
Compare full packagesTwo A40 lots represent two 48 GB GPUs each and USD 136.00 for 7 days, or USD 68.00 per lot. Choose this composition when each variant fits on one card and your software is compatible: the quantity and budget correspond directly to the two experiments you are running.
View the A40 and its plansWe allow direct payment of the plan in crypto. You do not have to top up an additional balance to pay for a single campaign. Keep this path when you want to apply your payment to the selected order; top-ups remain available for your subsequent rentals.
Choose direct payment or balanceNVIDIA RTX A5000 24GB
USD 23.571 GPUs total · 24 GB per card
Lot price: USD 23.57
Spend three days on a first trial: baseline run, memory measurement and saving results, with your prepared protocol.
1 lot to configure.
NVIDIA A40 48GB
USD 136.002 GPUs total · 48 GB per card
Lot price: USD 68.00
Assign a 48 GB card to each variant. Work on the same corpus and the same criteria, with a shared budget for both experiments.
Choose 2 lots in the configurator.
NVIDIA A100 SXM 80GB
USD 990.001 GPUs total · 80 GB per card
Lot price: USD 990.00
Keep an 80 GB card for thirty days to organize your test series, resume your experiments and gather their results.
1 lot to configure.
Each scenario corresponds to its own duration and quantity. These totals serve to choose a work schedule, not to rank GPU speed.
Keep the same GPU and the same duration. Choose one batch to run your trials in sequence, or two batches to assign a card to each independent variant.
| Your schedule | Rented capacity | Budget for 7 days | Action |
|---|---|---|---|
| Variants run successively | 1 batch · 1 48 GB GPU | USD 68.00 | Configure 1 batch |
| Two independent variants | 2 batches · 2 48 GB GPUs | USD 136.00 | Choose 2 batches in the configurator |
The link opens A40 over 7 days: choose 1 or 2 batches in the configurator to see the total for your scenario.
Configure A40 · 7 days| Configuration | Memory per GPU | GPU / lot | Available lots | 3 days · USD | 7 days · USD | 30 days · USD |
|---|---|---|---|---|---|---|
| NVIDIA RTX A5000 24GB | 24 Go | 1 | 51 | USD 23.57 | USD 55.00 | USD 190.00 |
| NVIDIA GeForce RTX 3090 24GB | 24 Go | 1 | 186 | USD 27.86 | USD 65.00 | USD 230.00 |
| NVIDIA A40 48GB | 48 Go | 1 | 37 | USD 29.14 | USD 68.00 | USD 240.00 |
| NVIDIA L4 24GB | 24 Go | 1 | 104 | USD 30.00 | USD 70.00 | USD 250.00 |
| NVIDIA RTX A6000 48GB | 48 Go | 1 | 65 | USD 35.14 | USD 82.00 | USD 290.00 |
| NVIDIA GeForce RTX 4090 24GB | 24 Go | 1 | 265 | USD 47.14 | USD 110.00 | USD 390.00 |
| NVIDIA RTX 6000 Ada 48GB | 48 Go | 1 | 26 | USD 55.71 | USD 130.00 | USD 460.00 |
| NVIDIA L40S 48GB | 48 Go | 1 | 74 | USD 63.43 | USD 148.00 | USD 520.00 |
| NVIDIA GeForce RTX 5090 32GB | 32 Go | 1 | 100 | USD 66.43 | USD 155.00 | USD 540.00 |
| NVIDIA A100 SXM 80GB | 80 Go | 1 | 78 | USD 120.00 | USD 280.00 | USD 990.00 |
| NVIDIA H100 PCIe 80GB | 80 Go | 1 | 27 | USD 180.00 | USD 420.00 | USD 1,490.00 |
| AMD Instinct MI300X 192GB | 192 Go | 1 | 8 | USD 192.86 | USD 450.00 | USD 1,590.00 |
| NVIDIA H100 SXM 80GB | 80 Go | 1 | 63 | USD 203.57 | USD 475.00 | USD 1,690.00 |
| NVIDIA H200 SXM 141GB | 141 Go | 1 | 19 | USD 274.29 | USD 640.00 | USD 2,290.00 |
| 2 × NVIDIA B200 SXM | 180 Go | 2 | 4 | USD 887.57 | USD 2,071.00 | USD 7,391.00 |
A B200 lot already includes its two cards. The prices in the table cover the entire period shown.
We build your amount around three choices: the GPU model, the duration and the number of lots. Each price covers one lot for the entire period. To get the rental budget, multiply this price by your quantity of lots, once.
A lot refers to the cards included in that price: one card for each configuration, two cards for B200. Two B200 lots therefore represent four GPUs and two plans, with no additional multiplication by the number of cards. This way you compare explicit quantities, even when the composition of the offers differs.
| Your choice | What the amount covers |
|---|---|
| Period | Full plan of 3, 7 or 30 days; minimum of 3 days, with no automatic renewal |
| Currency | USD for the total rental; crypto amount specified in the payment instructions |
| Price unit | 1 lot for the entire chosen period |
| Quantity | From 1 to 10 lots; rental total = plan price × number of lots |
| Cards per lot | 1 GPU, except B200: 2 GPUs already included in the lot price |
| First plan | USD 23.57 for 1 lot of NVIDIA RTX A5000 24GB, i.e. 1 GPU with 24 GB per card, for 3 days |
| Payment | Direct crypto payment with no mandatory prior top-up |
| Account, no KYC | First name, last name, email and password required; no ID document or KYC procedure requested in the IteraGPU path |
For a targeted validation whose data and steps are ready, we recommend starting with 3 days. Choose 7 days to organize a series of variants and their evaluation, or 30 days for a campaign with several stages. Include resumptions, checks and result export in your schedule.
If you are preparing two independent variants compatible with 48 GB per card, choose two A40 lots over 7 days: USD 136.00 for two GPUs, at USD 68.00 per lot. If your runs can follow one another on a single card, compare this budget with that of one lot over the necessary period. The right choice depends on your program and the runtime of your runs.
First select the configurations that match the memory and backend of your workload, then compare their plans over the same duration. Our 3, 7 and 30 day scenarios illustrate different programs: they serve to help you choose your commitment, without ranking the cards by speed.
You can pay for your order directly, in particular in BTC on Bitcoin or in USDT on Tron, or use a sufficient USD balance. The IteraGPU path requires neither an ID document nor a KYC procedure; your account with first name, last name, email and password keeps the tracking. See the payment page for the eight asset/network pairs and the top-up option.
The plan price sets the rental amount in USD. The payment instructions specify the crypto amount and its due date; your wallet or sending platform may apply its own network or withdrawal fees. Add these fees to your forecast when they apply, without confusing them with the GPU rate.
Before committing to a plan, have the elements that shape your campaign confirmed: software environment, required server resources, and access timeline. The ordering terms also cover the tax or refund points that matter for your purchase. Displayed stock does not reserve a future date, and payment confirmation comes before access is prepared.
Before ordering, specify the software, RAM, storage and start date your campaign requires. We invite you to confirm these requirements with the delivery terms.
Payment confirmation comes before the GPU is made available. Catalog stock counts lots; an order does not automatically reserve a future date.
The USD plan is separate from any network or withdrawal fees charged by your wallet or platform. See the order terms for the applicable conditions.
At IteraGPU, the first plan costs USD 23.57 for 1 lot of NVIDIA RTX A5000 24GB, i.e. 1 GPUs with 24 GB per card for 3 days. The amount covers that entire period and is paid directly in crypto, with no mandatory top-up. The process involves no KYC or identity document; first name, last name, email and password are required for the account. Any fees charged by your wallet remain separate from the rental price.
Choose an IteraGPU plan when your campaign has a defined working window: you commit the full price for 3, 7 or 30 days, with no automatic renewal. That period lets you organize computations, evaluations and exports within the same rental budget. The plan does not depend on the number of compute hours actually used; if you want to pay for only a few hours, this purchase model does not fit that criterion.
At IteraGPU, a lot contains 1 GPU for all configurations except B200, whose lot includes 2 GPUs of 180 GB each. Two A40 lots give two cards; two B200 lots give four cards. The total remains the price per lot for the chosen duration multiplied by the number of lots. The two memories in a B200 lot do not automatically become a single usable space.
We recommend two A40 IteraGPU lots when each variant fits on a 48 GB card and your software environment is compatible. For 7 days, you commit USD 136.00 in total for 2 lots and 2 GPUs, i.e. USD 68.00 per lot. Select 2 lots in the configurator. This setup gives you one card per experiment; the number of completed runs depends on your workload.
The IteraGPU price states the GPU plan amount in USD. The instructions give the crypto amount to send, its network and its deadline. Your wallet or platform may charge separately for a transfer or withdrawal: check its summary before sending. The absence of these fees from the GPU table does not mean they are free or included in the rental amount.
For an ML program whose memory requirements and working window you already know, choose our plans: the model, lots, period and rental amount are explicit before ordering. Pick your scenario, then configure the capacity you need and pay directly in crypto.
USD 23.57 for 1 lot of NVIDIA RTX A5000 24GB · 1 GPU with 24 GB · 3 days.