Oblivus Cloud rents H100, A100 and RTX GPUs with per-minute billing and free data transfer. GPU rates from $0.55/hr on demand. See features and pricing.
Researched and fact-checked by Bartu Cavusoglu · Updated October 26, 2026
Oblivus Cloud is a cloud platform that rents GPU and CPU instances for compute-heavy work. The company says it runs 9 data centers and more than 18,000 GPUs across 8 GPU SKUs, and it claims prices up to 80% below larger cloud providers.
Its homepage describes use cases including deep learning and AI, VFX, cloud gaming and institutional research computing. Instances are available on demand or through reservations, and the company says accounts carry no quotas or complicated verification steps.
Users create an account, add funds and deploy an instance from the console or the API. Billing is calculated per minute for active instances, and the balance is deducted in real time while an instance runs. Billing stops immediately when it is terminated.
Once an instance is deployed, its price and discount are locked in. If Oblivus lowers prices later, the customer benefits, but an increase does not change the locked rate. The model is prepaid, and Auto Top-Up can turn it into pay-as-you-go.
The GPU catalog covers H100, A100 80GB, L40, RTX 4090 and RTX A6000, with 1, 2, 4 and 8 GPU configurations on most models. Prices include vCPUs, RAM and NVMe storage. CPU-only Intel Xeon instances are priced per vCPU. Storage is replicated three times.
The API has more than 30 endpoints for deploying, modifying and scaling instances. Data ingress and egress are listed as free, and 24/7 support is included. Teams can create organizations, invite members by email and assign roles that control resource sharing and billing access.
Oblivus Cloud is a cloud platform that rents GPU and CPU instances on demand or by reservation. The company says it runs 9 data centers and more than 18,000 GPUs. It is aimed at AI, VFX, gaming and research workloads.
On-demand GPU prices on the pricing page run from $0.55/hr for an RTX A6000 to $4.31/hr for an 8x H100 SXM5 configuration. CPU instances start at $0.029/hr. Reservations are cheaper, with RTX A6000 from $0.38/hr, and final reservation pricing depends on term and contract size.
No. The company states it does not offer traditional free trials. You can test the platform with as little as $5, and occasional promotional credits may be offered.
Billing is calculated per minute for active instances, and the balance is deducted every minute while an instance runs. Billing stops immediately on termination. The model is prepaid, and Auto Top-Up can refill the balance automatically.
The pricing page lists H100 (SXM5, NVLINK and PCIe), A100 80GB (NVLINK and PCIe), L40, RTX 4090 and RTX A6000. Most come in 1, 2, 4 and 8 GPU configurations. The homepage also announces RTX 5090s on demand.
Yes. The API has more than 30 endpoints for deploying, modifying and scaling instances. Your API key is in the console under API, then Credentials, and compromised keys can be reset instantly.
Oblivus accepts credit and debit cards, Oblivus Pay, and wire transfers for business accounts. The pricing page also mentions bank transfers through Revolut and cryptocurrency payments through Oblivus Pay.
The company says personal accounts suit individual developers and small projects. Teams can create organizations with invited members and roles. Business accounts add priority support with an 8-hour response guarantee on high-priority tickets.