TheAI Cloud
- Free allowance
- No free tier and no single-GPU rental: whole 8-GPU bare-metal nodes only.
- Verdict
- paidno free allowance; listed for its price
- Cost
- no price stated
- Identity
- institutionno card, no phone
- Binding limit
- write-gatedlaunching instances requires prior verification (know-your-customer and export-control screening); 36 and 60 month terms are take-or-pay with 30 percent at signing
- Evidence
- documented4 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 and no single-GPU rental: whole 8-GPU bare-metal nodes only. B200 in Los Angeles USD 8.40 per GPU-hour month-to-month from November 2026; B300 in Kuala Lumpur USD 9.00 for 3 months paid up front; B300 in Iowa USD 4.70 on 36 months and USD 4.50 on 60. No H100, A100, L40S, RTX 4090 or L4 is listed. unit none; metered on wall-clock; none |
|---|---|
| Price | none stated |
| Limits |
Launching needs KYC and a signed capacity agreement; the nodes are sold whole (8 GPUs) and the cheapest terms are 36 or 60 months take-or-pay. |
| Identity | institution The terms state the service 'is offered to businesses and professional users only' and that 'Launching instances requires prior verification (know-your-customer and export-control screening)'. The order path is a quote request, KYC, then a master services agreement and an invoice. Payment is by wire transfer, card or USDC per the index; the vendor page names wire transfer for the 30 percent prepayment. |
| Egress | not published on the pages read |
| Interface | ssh, rest inferred from the tooling the provider documents; not probed |
| Tooling | SSH with root onto whole bare-metal nodes with the organisation's public keys installed; a machine-readable rate card at GET /api/v1/pricing. |
| Composes with | A Blackwell-only lane: none of the five common cards is sold. The rate is competitive for B200 and B300 but needs KYC, a quote and a multi-month term. |
| Reliability | The index notes the service was not available until Q4 2026 and is limited to a US data centre; the vendor page states B200 is available from November 2026 and states a 99.975 percent monthly uptime SLA for the Kuala Lumpur block. No outage was observed. |
| Notes | Found via computeprices (listed as TheAI Cloud, slug cloud-theai). Operator TheAI LTD, DIFC. The index lists 25 configurations in its header but one B200 row in the table. |
| Parsed numbers |
|
Evidence 4
| date | source | what it proved |
|---|---|---|
| 2026-11-03 | computeprices.com/providers/cloud-theai | the index lists only the B200 (8 GPUs, Los Angeles) at USD 8.40 per GPU-hour; no H100, A100, L40S, RTX 4090 or L4 |
| 2026-11-03 | cloud.theai.com/pricing | first-party rate for the B200: USD 8.40 per GPU-hour, whole 8-GPU nodes, billed monthly |
| 2026-11-03 | cloud.theai.com/terms | business-only service; launching instances needs KYC and export-control screening |
| 2026-11-03 | cloud.theai.com/pricing | the order path is a quote request, then KYC, then a master services agreement and an invoice |
Unknown values 3
- egress metered: egress metering was not published
- per object cap stated: the field does not apply to this kind of offer
- retention stated: the field does not apply to this kind of offer