Dataoorts Pricing
Dataoorts offers pricing that is 10x times lower than that of public cloud providers. Our goal is to make computational power affordable and accessible to everyone.
Fully Dedicated GPU VMs and Bare Metal Servers โ No Containers.
Introducing our Latest
โX-Series GPU VMsโ
Dataoorts X-Series Instances: Optimized for Compute-Intensive Workloads
- Ideal for AI & ML: Perfect for tasks such as fine-tuning and training large AI models.
- Powered by DDRA Cluster: Our cutting-edge Super DDRA Cluster ensures top-tier performance for compute-heavy tasks.
- No Containers, Just Power: Our instances are true KVM-based VMs or bare metal servers. You get dedicated, unshared GPU power.
- Dynamic Pricing: Costs vary based on real-time users, compute workloads, region, and other usage factors.
- On-Demand Availability: Operates seamlessly on the Super DDRA cluster, available when needed.
X-Series Instance Pricing
Typically $0.3 to $2.7 per hour
Pricing is dynamically allocated per second. For the most current rates, please login to the console.
- We dynamically manage instance resources such as RAM, CPU, and more. All Instances operate on DDRA Technology.
- X-Series Instances start with
x_, such asx_A100x2. This example indicates an instance with an A100 GPU Model, Total 160 GB of VRAM, & double GPU Configuration. - X-Series Instances are hosted in Tier 3 and 4 Data Centers, Ensuring maximum security and adherence to Zero-Trust Policies.
- Full Administrative/Root Access to Virtual Machine.
โStop your VM when not in use, run it only when needed โ save costs, stay efficient.โ
Stop Instance to Save Costs
When stopped, you are only charged $0.036 per hour.
Stop your instance to save costs.
- Currently, only X-Series instances support the stop/resume feature. Support for Spot and GC2 instances will be added soon.
- Stopping an instance immediately halts regular billing. However, please allow 20 Mins. for a snapshot checkpoint to be created.
- It's recommended to wait 20 Mins. after stopping the VM before restarting it to ensure the snapshot-checkpoint has completed.
- The public IP address remains the same and is reserved while the instance is stopped.
- The maximum size of the persistent snapshot disk is 95 GiB.
GPU Marketplace Offers
Secure reserved GPU instances at highly affordable rates through flexible bidding. Choose from short-term commitments starting at just one week to long-term reservations up to three years, ensuring cost-effective, reliable GPU VMs tailored to your needs.
Explore Marketplace
Super affordable GPU compute is now a realityโwith us!
Loading comparison data...
FAQs
Yes, you can create a persistent volume in a specific region and mount it to your GPU VM for continuous storage. This way, you wonโt need to re-download your data each time. The cost is $0.00059 per GB per hour.
Yes, you can. However, when you stop the instance, please allow 15โ20 minutes for a snapshot checkpoint to be created. Once the snapshot is completed, you can safely restart the instance. Do not interrupt the process while the snapshot is being generated.
Spot GPU instances (labeled with โ-spotโ) are idle resources offered at a lower cost. When demand arises, they may be reclaimed for other workloads. Theyโre ideal for serverless or sandbox tasks, as they provide a more affordable option compared to on-demand instances.
Yes GPU instances can be stopped. Once stopped, they incur a minimal charge of $0.036 per hour.
Please note: Only the root disk is persistent, and the VM state is preserved when restarting from a stopped instance. Any ephemeral data will be lost after a restart.
X-Series Instances dynamically adjust pricing every second within a specified range.
Learn About DDRA: Here
Our new X-Series Instances, operating on our new advanced DDRA cluster, Are optimized for compute-intensive workloads, making them ideal for tasks such as fine-tuning and training large AI models.
X-Series Instances operate on our advanced DDRA Cluster, designed for heavy compute-intensive workloads. The cost of X-Series Instances varies within the range displayed on our dashboard and depends on several factors, including the total number of real-time users on the cluster, ongoing compute workloads, regions where instances are in use, and more.
