Description
What you get
- 1× NVIDIA A100, 40 GB VRAM — dedicated passthrough for the life of your subscription
- 16 vCPUs · 128 GB RAM
- 2 TB local NVMe
- Fully customisable environment with root access
- Priority chat and email support
- 99.9% uptime SLA
Who this is for
Teams training and running models in production. The 40 GB of VRAM is what you want for LoRA on 13B and above, full fine-tuning of smaller models, long-context work where activation memory dominates, and larger batches where gradient accumulation is costing you real time.
It is also the plan to choose for anything you intend to run for weeks. Sitting close to the memory ceiling means one unusually long sample can kill a run you have been babysitting.
Dedicated means dedicated
This is a whole physical card mapped to your instance — no MIG partition, no time-slicing, no preemption. You can verify it yourself with nvidia-smi, and we explain exactly how to check. The commitment is written into Terms §3 rather than left to a sales conversation.
Billing
$599 per month in USD, exclusive of tax. No setup fee, no lock-in. Your first month is covered by a 14-day full refund. Instances run on our own hardware in Mumbai, India.
How to order
A100 capacity is finite, so we confirm availability before taking payment. Send an enquiry and we will reply within one business day.
Reviews
There are no reviews yet.