- Free allowance
- No free tier. Prepaid per-second GPU rental in two tiers.
- Verdict
- paidno free allowance; listed for its price
- Cost
- no price stated
- Identity
- email?placeholder, not evidence; card unknown, phone unknown
- Binding limit
- noneno named constraint binds
- Evidence
- documented12 sources; a first-party page was read and the sentence quoted
- Freshness
- stale on 2026-12-04checked 2026-11-03 (live); stale on 2026-12-04
| Free tier | No free tier. Prepaid per-second GPU rental in two tiers. Standard tier (no enclave): from USD 0.32 per GPU-hour; floors of USD 0.70 for the A100 80 GB and USD 0.61 for the RTX 4090 on the vendor page, and the index lists the H100 SXM at USD 3.60, the L40S at USD 0.76 and the RTX 4090 at USD 0.70. Confidential VM tier (Intel TDX): H100 USD 6.95, H200 USD 8.08, RTX PRO 6000 Blackwell USD 3.80 per GPU-hour. The index lists no L4. The only allowance is one RTX PRO 6000 hour (USD 3.80) added to a first top-up of USD 10 or more. unit none; metered on wall-clock; none |
|---|
| Price | none stated |
|---|
| Binding | none No quota is stated: the wall is the prepaid balance and the per-GPU-hour rate. Capacity is first come, first served and the standard tier is a distributed host network with no enclave. |
|---|
| Identity | email (placeholder) unverified: the signup page answered with a login wall. The terms require 'accurate, complete, and current information during registration', and billable use needs a prepaid balance of at least USD 5 paid by card (Stripe), cryptocurrency (Coinbase Commerce) or Bitcoin, so a card is optional. A phone or ID requirement is not stated on the pages read. |
|---|
| Egress | not published on the pages read |
|---|
| Interface | rest, cli, ssh inferred from the tooling the provider documents; not probed |
|---|
| Tooling | REST API at /api/volt/pods with an X-API-Key header, the voltagegpu CLI (pip install voltagegpu-cli) and SSH with root into each machine; templates for PyTorch, Jupyter and Ollama. |
|---|
| Composes with | Per-second billing with a USD 5 minimum top-up makes short GPU runs cheap to start: an L40S at USD 0.76 or a 4090 at USD 0.61 to USD 0.70 an hour. Capacity is first come, first served, with no reservations. |
|---|
| Reliability | The index records 'Prepaid only, no invoicing on credit' and 'No reservations: capacity is first come, first served'; on 8x H100 confidential nodes the NVSwitch fabric is not attested. No outage was observed. |
|---|
| Notes | Found via computeprices (listed there as VoltageGPU). The entity is VOLTAGE EI, France. The index prices the H100 at USD 3.60 on the standard tier, and the vendor page shows USD 6.95 for the confidential tier; the vendor page lists no standard-tier H100 row. A100 and L4 are not on the index page; the A100 80 GB floor is the vendor's. |
|---|
| Parsed numbers | |
|---|