Deep Learning optimized VMs

Linux Dedicated Servers with 1 to 8 NVIDIA Ampere A6000 48 GB GPUs together with up to 48 Core Xeon CPU, 32 to 768 GB of memory and fast SSD storage.

SIGN IN

100 Hours of GPU Compute available for free for eligible organizations. Sign In to find out how you can benefit.

Ubuntu or CentOS Dedicated Servers

  • 16 – 48 Core Xeon Gold CPUs
  • 1 – 8 RTX A6000 48 GB Professional GPUs
  • 32 – 512 GB ECC Memory
  • 100 GB Operating System Disk
  • Up to 10 TB Storage
  • Installation and Setup support
  • Simple Pricing

Virtual Machines with NVIDIA Professional Cards

Billed per Hour, Week or Month. Available for Organizations and Companies.

A 6000
$ 0.8
Per Hour
1 Nvidia Ampere A6000
48 GB ECC Video Memory
12 vCPU Cores?
48 GB RAM?
120 GB OS Disk
Up to 4 TB Storage?
1 TB Bandwidth
100 Hours Free?
SIGN IN
R 6000
$ 0.6
Per Hour
1 Nvidia Quadro RTX 6000
24 GB ECC Video Memory
12 vCPU Cores?
24 GB RAM?
120 GB OS Disk
Up to 4 TB Storage?
1 TB Bandwidth
100 Hours Free?
SIGN IN
A 6000
$ 129
Per Week
1 Nvidia Ampere A6000
48 GB ECC Video Memory
24 vCPU Cores?
48 GB RAM?
120 GB OS Disk
Up to 4 TB Storage?
1 TB Bandwidth
100 Hours Free?
SIGN IN
R 6000
$ 99
Per Week
1 Nvidia Quadro RTX 6000
24 GB ECC Video Memory
24 vCPU Cores?
24 GB RAM?
120 GB OS Disk
Up to 4 TB Storage?
1 TB Bandwidth
100 Hours Free?
SIGN IN
A 6000
$ 499
Per Month
1 Nvidia Ampere A6000
48 GB ECC Video Memory
24 vCPU Cores?
96 GB RAM?
120 GB OS Disk
Up to 4 TB Storage?
1 TB Bandwidth
100 Hours Free?
SIGN IN
R 6000
$ 399
Per Month
1 Nvidia Quadro RTX 6000
24 GB ECC Video Memory
24 vCPU Cores?
48 GB RAM?
120 GB OS Disk
Up to 4 TB Storage?
1 TB Bandwidth
100 Hours Free?
SIGN IN

Configuration with 1, 2, 4 or 8 cards are available. Pricing is a multiple of the base price regardless of the GPU type. Ubuntu, CentOS or Windows VMs.

Hourly Billing is Coming Soon, only weekly and monthly commitments are available at the moment.

Deep Learning MLOps

Use any model on a VM you completely control. Develop limitless features for your model. Load the model in memory and use up to 8 GPUs with no limitations.

Optimal price-performance for GPU workloads

Our GPU compute A Instances are up to 30% cheaper than alternatives on AWS or Google Cloud, while providing more resources, unlimited bandwidth and  storage included.

ProviderPlanTFlopsCuda CoresGPUsGPU MemGPUGPU architectureHourlyMonthly
AWSp2.xlarge4.112496112GB GDDR5NVIDIA® Tesla® K80Kepler (2014)$0.9000/h$648/mo
Googlec2-standard-16 + NVIDIA® K80 GPUs4.112496112GB GDDR5NVIDIA® Tesla® K80Kepler (2014)$1.2852/h$925/mo
Auxilio1x.a600016.314608148GB GDDR6NVIDIA Quadro RTX 6000Ampere (2021)$0.8/h$499/mo
Auxiliop2.8xlarge4.11 x82496x8812GB GDDR5 x8NVIDIA® Tesla® K80Kepler (2014)$7.200/h$5184/mo
Googlec2-standard-60 + 8x NVIDIA® K80 GPUs4.11 x82496x8812GB GDDR5 x8NVIDIA® Tesla® K80Kepler (2014)$6.7321/h$4847/mo
Auxilio8x.a600016.31 x8 4608 x8848GB GDDR6 x8NVIDIA Quadro RTX 6000Ampere (2021)$6.4/h$3992/mo

FAQ

How am I charged?

The Virtual Machines have a flat rate per month. For the moment we can offer a great deal for the A6000 cards, of 129.99 per week or 499.99 per month. Hourly billing is coming soon and will only bill powered on VMs. Current offer is subject to availability.

All our GPU servers are hosted in Germany and Romania in professional grade Tier 3 data centers. There are multiple 20 Gb up-links available and data is stored on redundant flash drives.

We do not offer a free trial, but you can request a full refund if you’re experiencing technical issues. You’ll receive a refund within 5 – 10 business days. We do offer 100 Hours of GPU compute for free on available infrastructure.

Send a cancellation request in your account to fully cancel the service. You will not be billed from the next month on and until them you can download any data you want from your machines – models, your data, all saved games, etc.

Short answer is yes, get in touch and let’s see what are your organizations needs and can we help. We support educational, academic and research groups with cloud resources.